AtlasWork, planned itself.

    The AI-native, all-in-one work platform. Tasks, projects, CRM, contracts, and analytics in one calm workspace.

    System status
    • SSO
    • SCIM
    • Two-factor sign-in
    • Audit log

    Product

    • Overview
    • PDF tools
    • Diagram tools
    • People & HR
    • Integrations
    • Marketplace
    • Pricing

    Resources

    • Guides
    • Glossary
    • Compare
    • Docs
    • API reference
    • Support
    • Changelog
    • Status

    Company

    • About
    • Careers
    • Press
    • Contact

    Legal & trust

    • Trust center
    • Security
    • Privacy
    • Terms
    • DPA
    • GDPR
    • SLA
    • Refunds
    • Google API data
    Atlas, a product by wrxstack.com·© 2026 wrxstack·All rights reserved
    PrivacyTermsSecurityStatus

    You are in control of your cookies

    Atlas uses strictly necessary cookies to keep you signed in. With your consent, we add anonymized product analytics, conversion attribution, and remembered preferences. Change your mind any time at /privacy/cookies.

    Off until you agree · Change it any time

    • Necessaryalways on
    • Analyticsopt-in
    • Marketingopt-in
    • Preferencesopt-in
    Skip to documentation
    Docs
    Back to Atlas

    Start here

    • Overview

    Developer

    • REST API guide
    • Authentication
    • API reference
    • MCP (AI agents)
    • MCP tools reference
    • SDKs
    • Quick actions
    • Changelog

    Webhooks

    • Overview
    • Quickstart
    • Events
    • Payloads and headers
    • Security and signing
    • Delivery and retries
    • Managing via API

    Connect

    • Connectors
    • Integrations

    Product

    • Collaboration and chat
    • Signing in and security
    • Client portal

    Reference

    • Glossary
    • Keyboard shortcuts
    • Module reference

    Model Context Protocol

    Connect AI agents to Atlas with MCP

    Atlas ships an enterprise-grade Model Context Protocol server: 2326 tools, 11 prompt templates, structured logging, retry-with-backoff, request timeout, graceful shutdown, and in-process metrics. Connect Claude, ChatGPT, Cursor, VS Code, Claude Code and any other MCP host by URL, with nothing to install.

    Create an API token
    Live from /.well-known/atlas-mcp.json@atlas/mcp-server@0.3.0/2326 tools/11 prompts/transports: stdio, sse

    Connected in a minute, nothing to install

    Give your assistant the Atlas MCP address and sign in to Atlas when it asks. Assistants that cannot sign in take a Personal Access Token instead. To run the server on your own machine, download one file and run it with Node: no npm, no package install. Connect by URL.

    What is MCP?

    Model Context Protocol is an open standard for connecting AI assistants to external tools and data. The host (Claude, Cursor, Cline, Continue) speaks JSON-RPC over stdio or SSE; servers like Atlas expose a typed catalogue of tool calls the model can invoke. The same Atlas server works in every host: set it up once per device and you get it everywhere.


    Core concepts

    MCP servers expose three kinds of capability: tools that act, resources a host can attach to a conversation, and prompts that package a workflow. Atlas offers all three.

    Tools

    Typed actions the model can call, each with a name, a required scope, and a JSON input. Creating a task, reading a project, running a report. This is where almost all of the work happens. The full catalogue is below.

    Resources

    Records the host can attach to a conversation without calling a tool: you, a task, a project, a CRM account, a wiki page, or a service incident. Each is read through the same API route and with the same token as a tool, so it is exactly what the API returns. The addresses are listed further down this page.

    Prompts

    Reusable multi-step recipes the host surfaces in its prompt picker. Pick one and the host inserts the whole workflow for you. Atlas ships several out of the box, listed further down this page.

    How a tool call works

    You never invoke a tool by hand. You describe what you want in plain language, the host model picks the matching tool, fills its input from your request, and calls it over JSON-RPC. Atlas checks the PAT scope, runs the underlying REST call, and returns a typed result the model reads back to you.

    Say "Add a task: review the Q3 forecast, due Friday, in the Atlas project." The host resolves that to the atlas_tasks_create tool and sends an input like this:

    json
    {
      "tool": "atlas_tasks_create",
      "input": {
        "title": "Review the Q3 forecast",
        "projectId": "prj_atlas",
        "dueDate": "2026-07-10",
        "priority": "HIGH"
      }
    }

    Atlas returns the created task:

    json
    {
      "id": "tsk_a1b2c3",
      "title": "Review the Q3 forecast",
      "status": "todo",
      "priority": "HIGH",
      "dueDate": "2026-07-10",
      "projectId": "prj_atlas",
      "url": "https://app.atlas.example/tasks/tsk_a1b2c3"
    }

    Exact fields per tool follow the REST contract. See the REST API reference for the full request and response schema behind each tool.


    Two ways to connect

    Connect by URL for a connection with nothing to install, or run the server on your own machine from one downloaded file. Both offer the same tools.

    Recommended

    No install

    Connect by URL

    Paste the Atlas MCP address into your assistant and sign in to Atlas, or send a Personal Access Token. Nothing to install.

    • Sign in with Atlas: no token to copy or store
    • Works with Claude, ChatGPT, Cursor, VS Code and Claude Code
    • End access any time in Settings, Connected apps
    Connect by URL

    One file, no npm

    Run it on your machine

    Download one file from Atlas, check its checksum, and run it with Node. No npm and no package install.

    • For a token limited to known IP addresses
    • Works behind networks that block outside connections
    • Runs in any assistant that starts a local command
    Run it on your machine

    Not sure? Connect by URL. Run it on your own machine only when your token is limited to known IP addresses, your network blocks outside connections, or a policy asks for it.


    Connect by URL (recommended)

    Give your assistant the Atlas MCP address. Assistants that sign in open the Atlas consent screen; others send a Personal Access Token. There is nothing to install.

    Atlas MCP address

    https://api-atlas.wrxstack.com/mcp

    An assistant that supports sign-in registers itself with Atlas, opens the Atlas sign-in and consent screen, and refreshes its access on its own. An assistant that takes a header sends a Personal Access Token as Authorization: Bearer atlas_pat_.... Either way the assistant acts as you and sees only what you can see.

    Claude (claude.ai, Claude Desktop and the mobile apps)

    1. 1

      Add Atlas as a custom connector

      Open Settings, then Connectors, and add a custom connector. Name it Atlas and paste the Atlas MCP address above.

    2. 2

      Sign in and approve

      Connect. Sign in to Atlas in the window that opens, review what Claude is asking to do, and approve. The Atlas tools appear in Claude.

    Custom connectors are available on the Claude plans that offer them; on a Team or Enterprise plan an owner may need to add the connector for the organization first.

    ChatGPT

    Where your ChatGPT plan supports custom connectors, add one in the connector settings with the Atlas MCP address and choose OAuth as its authentication. ChatGPT opens the Atlas sign-in and consent screen; approve to finish.

    Claude Code

    Add the server, then run /mcp in Claude Code and choose to authenticate Atlas:

    bash
    claude mcp add --transport http atlas https://api-atlas.wrxstack.com/mcp

    VS Code

    Add this to .vscode/mcp.json in your project (or to your user MCP configuration). VS Code asks you to sign in to Atlas the first time Atlas is used.

    json
    {
      "servers": {
        "atlas": {
          "type": "http",
          "url": "https://api-atlas.wrxstack.com/mcp"
        }
      }
    }

    Cursor

    Add this to ~/.cursor/mcp.json (or your project's .cursor/mcp.json) and restart Cursor. Cursor asks you to sign in to Atlas.

    json
    {
      "mcpServers": {
        "atlas": {
          "url": "https://api-atlas.wrxstack.com/mcp"
        }
      }
    }

    Any client that sends a header

    A client that cannot sign in sends a Personal Access Token instead. Mint one in Settings, API access with the narrowest scopes that fit, and put it in the header:

    json
    {
      "mcpServers": {
        "atlas": {
          "url": "https://api-atlas.wrxstack.com/mcp",
          "headers": {
            "Authorization": "Bearer atlas_pat_REPLACE_WITH_YOUR_TOKEN"
          }
        }
      }
    }

    A client that only starts local commands

    Run the open-source mcp-remote bridge, which connects to the Atlas address for the client, or run the Atlas server on your machine.

    bash
    npx -y mcp-remote https://api-atlas.wrxstack.com/mcp \
      --header "Authorization: Bearer atlas_pat_REPLACE_WITH_YOUR_TOKEN"

    What you approve, and how to end it

    The consent screen names the assistant, says if it registered itself with Atlas, lists what it may do in plain words, and shows where you return afterwards. End an assistant's access at any time in Settings, Connected apps; a token stops working at once. Revoke a Personal Access Token in Settings, API access.

    Same security as the API

    Every request is checked against the same scopes, workspace isolation and rate limits as the Atlas API, and only tools for normal accounts are offered here. A Personal Access Token limited to known IP addresses cannot be used through the address, because Atlas makes the calls for you; run the server on your own machine for that token.

    Run it on your own machine (no npm)

    Download one file from Atlas and run it with Node 18.17 or later. Use it when your token is limited to known IP addresses, your network blocks outside connections, or a policy asks for a local server.

    1. 1

      Mint a Personal Access Token

      Open Settings, API access and create a token with the narrowest scopes that fit. Start with tasks:read and projects:read if the assistant only reads. The token starts with atlas_pat_ and is shown once, so copy it now.

    2. 2

      Download the server and check it

      The file and its SHA-256 checksum come from the Atlas API. On macOS or Linux, this saves the file and its checksum, checks it (the check must print OK), and prints the path to use:

      bash
      mkdir -p "$HOME/.atlas" && cd "$HOME/.atlas"
      curl -fsSLO https://api-atlas.wrxstack.com/mcp/download/atlas-mcp-server.mjs
      curl -fsSLO https://api-atlas.wrxstack.com/mcp/download/atlas-mcp-server.mjs.sha256
      sha256sum -c atlas-mcp-server.mjs.sha256 2>/dev/null || shasum -a 256 -c atlas-mcp-server.mjs.sha256
      echo "$HOME/.atlas/atlas-mcp-server.mjs"

      On Windows, in PowerShell, the two checksums printed must match:

      bash
      New-Item -ItemType Directory -Force "$HOME\.atlas" | Out-Null
      Invoke-WebRequest https://api-atlas.wrxstack.com/mcp/download/atlas-mcp-server.mjs -OutFile "$HOME\.atlas\atlas-mcp-server.mjs"
      (Get-FileHash "$HOME\.atlas\atlas-mcp-server.mjs" -Algorithm SHA256).Hash.ToLower()
      (Invoke-WebRequest https://api-atlas.wrxstack.com/mcp/download/atlas-mcp-server.mjs.sha256).Content
      "$HOME\.atlas\atlas-mcp-server.mjs"
    3. 3

      Configure your assistant

      Pick the host you use. Replace the token, and replace /Users/you/.atlas/atlas-mcp-server.mjs with the path the download printed. Then save and restart the host.

      Add this to the Claude Desktop config, replace the token and the path, then quit Claude Desktop completely and open it again (Cmd+Q on macOS; closing the window is not enough).

      macOS: ~/Library/Application Support/Claude/claude_desktop_config.json. Windows: %APPDATA%\Claude\claude_desktop_config.json

      json
      {
        "mcpServers": {
          "atlas": {
            "command": "node",
            "args": [
              "/Users/you/.atlas/atlas-mcp-server.mjs"
            ],
            "env": {
              "ATLAS_API_KEY": "atlas_pat_REPLACE_WITH_YOUR_TOKEN",
              "ATLAS_API_URL": "https://api-atlas.wrxstack.com"
            }
          }
        }
      }
    4. 4

      Try it

      After restarting the host, the Atlas tools appear in its tool list. Try one of these:

      • "Run the atlas_server_health_check tool." The token needs projects:read, because the tool checks access with /v1/projects?limit=1.
      • "List my open Atlas tasks."
      • "Create an Atlas task: review the Q3 forecast, due Friday."

    What each tool tells your host

    Before a call, every tool says whether it only reads, whether its change can be undone, and whether repeating it is safe. After a call, the result comes back as structured data as well as text, and a failure says what went wrong in the words of the API.

    HintWhat it means
    readOnlyHintTrue when the tool only reads. A tool that reads nothing but builds a link, such as a download address, is read only too.
    destructiveHintTrue when its change cannot be taken back: something deleted, a message sent, money charged, access revoked.
    idempotentHintTrue when calling it twice with the same input has the same effect as calling it once.
    openWorldHintAlways false. Every tool acts on your Atlas workspace and nothing outside it.

    Hosts use these to decide when to ask you first. Each tool also carries a title, a short name such as List tasks. The hints are advice to the host: the API still decides what your token may do.

    Every result comes back as structured content and, for hosts that read only text, as the same JSON in the text content. Structured content is always an object, so a list comes back as { items } and any other value as { value }.

    A failed call is marked as an error and keeps the problem the API returned, so the model can tell you what happened and what to do:

    json
    {
      "isError": true,
      "content": [
        {
          "type": "text",
          "text": "[API_ERROR] 404 Task not found. No task with id tsk_9 in this tenant (https://atlas.dev/errors/not-found, task-not-found)"
        }
      ],
      "structuredContent": {
        "error": {
          "type": "https://atlas.dev/errors/not-found",
          "code": "task-not-found",
          "title": "Task not found",
          "status": 404,
          "detail": "No task with id tsk_9 in this tenant"
        }
      }
    }

    A tool that reads a list returns one page as { items, nextCursor }. To read further, call the same tool again with cursor set to nextCursor, and stop when nextCursor is null. A cursor is opaque: pass it back exactly as it arrived. Each list tool names the limit, sort and order it accepts.

    json
    // First call
    { "name": "atlas_tasks_list", "arguments": { "projectId": "prj_42", "limit": 50 } }
    // It returns { "items": [ ... ], "nextCursor": "tsk_c9" }
    
    // Next call
    { "name": "atlas_tasks_list", "arguments": { "projectId": "prj_42", "limit": 50, "cursor": "tsk_c9" } }
    // Stop when nextCursor is null

    Resources

    Records a host can attach to a conversation. Each address is read through the API route shown, with your token, so a resource is exactly what that route returns and only what your token may see.

    AddressWhat it isRead through
    atlas://meYou: name, email, role, time zone and profile.GET /profile/me
    atlas://tasks/{id}One task with its labels.GET /v1/tasks/{id}
    atlas://projects/{id}One project.GET /projects/{id}
    atlas://crm/accounts/{id}One CRM account.GET /v1/crm/accounts/{id}
    atlas://wiki/pages/{id}One knowledge base page with its content.GET /v1/wiki/pages/{id}
    atlas://service/incidents/{id}One service incident with its response roles, updates and attached conversations.GET /v1/service-ops/incidents/{id}

    In each address, put the record id where the template shows id. The Super Admin server offers no resources.


    One-click prompt templates

    MCP-aware hosts surface these in their prompt picker. Pick one and the host inserts the multi-step workflow for you, no need to memorize the recipe.

    PromptWhat it does
    atlas_daily_standupDraft my standup for today: what I finished recently, what I am working on, and what is blocked.
    atlas_sprint_reviewReview a project sprint: what shipped, what slipped, the top risks, and a one-paragraph summary.
    atlas_backlog_groomingGroom a project backlog: suggest a priority for each unscheduled task and flag likely duplicates, without changing anything.
    atlas_inbox_zeroWork through my overdue tasks and the ones due today, asking one short question about each.
    atlas_create_project_with_templateSet up a new project: ask me for three to five milestones, then create the project and a task for each.
    atlas_crm_pipeline_reviewReview the sales pipeline: value by stage, deals that have stalled, and the deals most worth attention this week.
    atlas_hr_leave_triageTriage pending leave requests: who is asking, for when, whether they have the balance, and where requests overlap.
    atlas_incident_summarySummarise one service incident: what happened, its impact, the timeline, who is responding, and what happens next.
    atlas_weekly_delivery_statusWrite the weekly status of a client engagement: readiness, open risks and issues, and what the client needs to know.
    atlas_contract_renewal_reviewReview contracts coming up for renewal: which expire soon, which already have a renewal in progress, and which need a decision.
    atlas_inbox_triageTriage my notifications: group them by what needs action, what is for information, and what can be archived.

    Tools the server exposes (2326)

    Each tool is a typed JSON-RPC call. The host model decides when to call which tool; you do not need to memorize the names. Scopes are checked on every call; missing a scope returns the API 403 insufficient_scope verbatim.

    Reading a tool entry

    Each entry shows the tool name (the JSON-RPC method the model calls), the PAT scope it requires, what it does, and an example phrase you can say to trigger it. Scope a token to just what you see here and the agent gets exactly those tools, nothing more. Search by name, scope, or module, or jump straight to a module below.

    How tools are named

    Every tool is named atlas_<area>_<resource>_<verb>: the product area first, then the thing the tool acts on, and the action last, as in atlas_crm_accounts_create or atlas_hr_leave_requests_approve. When the resource is the area itself, it appears once, as in atlas_tasks_list. Generated route bridges, named atlas_route_ followed by the method and the path, are the exception.

    The tools were renamed to this pattern in place, with no aliases. If a saved prompt or a host configuration names a tool, update it: an old name now returns an Unknown tool error. The changelog of the MCP server package lists every old name beside its new one.

    2326

    Jump to a module

    /
    • Client Delivery463
    • Client Onboarding32
    • Tasks82
    • Projects112
    • Portfolios2
    • Governance199
    • Admin46
    • Access39
    • People36
    • Notifications21
    • Activity1
    • CRM133
    • Contracts60
    • Agreements35
    • PDF Studio89
    • HR Core50
    • HR Engagement21
    • HR Payroll40
    • HR Hiring34
    • HR Performance8
    • HR Admin42
    • Self-Service40
    • Personal Productivity40
    • Goals13
    • Focus4
    • Workload26
    • Themes5
    • Voice3
    • Meetings33
    • Booking25
    • Service Operations65
    • Data Export4
    • Search4
    • Runtime11
    • Docs8
    • Sync2
    • Knowledge Base23
    • Automations21
    • Bulk4
    • Compliance11
    • Billing6
    • Email16
    • Changelog18
    • Agents27
    • Workspaces7
    • Reports12
    • Insights10
    • Intelligence5
    • Comments4
    • Growth2
    • Connectors166
    • Privacy2
    • Status Page8
    • Calendar9
    • Forms6
    • Chat44
    • Huddles6
    • Boards26
    • Diagrams48
    • Whiteboards7
    • eSign10

    Read one engagement and the registers that describe it: the risk log, the deliverables, what is owed by the client, the plan and its critical path, the workstreams, the governed meetings, the open threads and the stakeholder map. The write surface is deliberately two tools, raising a change request and logging a risk, because both record a decision somebody has already taken. Approving a change, issuing a status report and signing anything off are absent on purpose, and the commercial position is not exposed at all.

    • atlas_delivery_vocabulary_getdelivery:read

      The vocabularies this workspace uses: the follow-up rules with their kind, subject and default lead time, and the kinds, statuses and channels a follow-up can carry. Read this before writing to the module, so the values you send are ones the workspace actually accepts rather than a copy that drifted a release ago.

      "What follow-up kinds and statuses does this workspace accept?"

    • atlas_delivery_engagements_listdelivery:read

      List engagements with optional state, account, partner and free-text filters. Each row carries the reference, client, lifecycle state, health rating, dates and contract value. Use this to answer "what are we delivering" or "which engagements are red".

      "Which engagements are amber or red right now?"

    • atlas_delivery_engagements_getdelivery:read

      Read one engagement: its reference, client, lifecycle state, the seven health ratings, dates, contract value and approved change value. The contract value and the approved change are separate figures; the total is their sum.

      "Show me the Northwind transformation engagement."

    • atlas_delivery_engagements_readiness_getdelivery:read

      Answer "why can this engagement not start yet". Returns every blocker at once: an acceptance decision not taken or expired, a screening check outstanding or lapsed, a legal instrument not executed, or a quality review still owed. An empty list means nothing is blocking it.

      "Can the Northwind engagement start yet?"

    • atlas_delivery_engagements_raid_items_listdelivery:read

      Read the risk, assumption, issue, dependency and decision register for one engagement, with scores and heat bands. Use this to answer "what is going wrong" or "what are the severe risks".

      "What are the severe risks on Northwind?"

    • atlas_delivery_engagements_deliverables_listdelivery:read

      Read the deliverable register for one engagement: what the firm owes the client, and which review gate each piece is sitting at. Use this to answer "what is late" or "what is waiting on the client".

      "What deliverables are late on Northwind?"

    • atlas_delivery_engagements_info_requests_listdelivery:read

      Read the information request list for one engagement: what has been asked of the client, what has come back, and what is overdue or blocking the plan. Use this to answer "what are we waiting on".

      "What are we waiting on the client for?"

    • atlas_delivery_engagements_plan_getdelivery:read

      Read the plan for one engagement: every row of the work breakdown structure with its dates, percent complete, critical-path flag and float. Use this to answer "what is on the critical path" or "what slipped".

      "What is on the critical path for Northwind?"

    • atlas_delivery_engagements_workstreams_listdelivery:read

      Read the workstreams of one engagement, each with its own traffic light and progress. Use this to answer "which part of the work is in trouble" rather than judging the engagement as a whole.

      "Which workstream is in trouble?"

    • atlas_delivery_engagements_meetings_listdelivery:read

      Read the governed meetings of one engagement, with quorum and the state of their minutes. Quorum has three states and they differ: not evaluated is not the same as not met.

      "Which minutes are still unapproved?"

    • atlas_delivery_engagements_threads_listdelivery:read

      Read the open questions, blockers and escalations on one engagement, including which side each is waiting on. Use this to answer "what is unanswered".

      "What questions are unanswered on Northwind?"

    • atlas_delivery_engagements_stakeholders_listdelivery:read

      Read the stakeholder map of one engagement: who matters on the client side, how they feel about the work, and who is overdue a conversation.

      "Who are we overdue a conversation with?"

    • atlas_delivery_engagements_change_requests_listdelivery:read

      Read the change register for one engagement: every change request with its type, status, schedule impact in days and cost impact.

      "What changes have been raised on Northwind?"

    • atlas_delivery_engagements_change_requests_drift_getdelivery:read

      Cumulative drift from approved changes, measured against the ORIGINAL baseline rather than the current one. This is the number that catches ten changes each comfortably under the approval threshold whose sum has moved the engagement thirty percent from what was signed.

      "How far has Northwind drifted from its baseline?"

    • atlas_delivery_engagements_status_reports_listdelivery:read

      Read the status report history for one engagement, with each report period, the seven health ratings and whether it was issued.

      "Show me the last status report for Northwind."

    • atlas_delivery_engagements_storyline_getdelivery:read

      How far the deck actually is. Returns each section with its status and a progress summary counting evidenced pages rather than a completion percentage: a page counts as evidenced only when it has an answer written on it and a completed analysis behind it.

      "What is the storyline for the Northwind deck?"

    • atlas_delivery_engagements_hypothesis_tree_getdelivery:read

      The hypothesis tree with its advisory violations. The violation worth acting on is a concluded leaf with no completed analysis behind it, which is a conclusion with no evidence.

      "Which hypotheses are still untested?"

    • atlas_delivery_engagements_commercials_getdelivery:read

      Read the live financial picture of one engagement, computed at the moment of asking: budget at completion, actual cost, earned and invoiced revenue, unbilled work in progress, margin, realisation, burn rate, runway and the estimate at completion. Use this to answer "how is this engagement doing financially". Pass `asOf` to see the picture at an earlier date.

      "How is the Northwind engagement doing financially?"

    • atlas_delivery_engagements_budgets_listdelivery:read

      Read the budgets of one engagement, at engagement, phase, workstream or deliverable level, each with its planned effort, planned cost, planned revenue, contingency and management reserve. Use this to answer "what did we plan to spend".

      "What budget did we set for each workstream on Northwind?"

    • atlas_delivery_engagements_commercials_burn_getdelivery:read

      Read the burn of one engagement: the budget at completion, actual cost to date, burn rate per day, the days of runway left at that rate, and percent complete. Use this to answer "when does the budget run out".

      "When does the Northwind budget run out at the current burn?"

    • atlas_delivery_engagements_commercials_profitability_getdelivery:read

      Read the profitability of one engagement: planned, earned and invoiced revenue, unbilled work in progress, over-billing, actual cost, margin percent, realisation percent, effective rate, and the estimate and variance at completion. Use this to answer "are we making money on this engagement".

      "What margin are we making on Northwind?"

    • atlas_delivery_engagements_financial_snapshots_listdelivery:read

      Read the frozen financial snapshots of one engagement, newest first and at most 200. A snapshot keeps the figures as they stood on a day, so a number quoted to a client stays quotable after a timesheet is corrected. Use this to answer "what did the numbers say at the last report".

      "What did the Northwind numbers say at the end of last month?"

    • atlas_delivery_engagements_metrics_getdelivery:read

      Read the metric catalogue of one engagement over a period: lock-up, the health score, governance, quality, delivery, people and commercial measures, the operating model metrics, and a list of what could not be measured. Use this for a full scorecard of one engagement rather than one figure.

      "Give me the full scorecard for Northwind over the last 90 days."

    • atlas_delivery_engagements_subcontractors_listdelivery:read

      Read the subcontractor register of one engagement: each party with its role, workstream, contract status, rates, dates, and whether insurance, the confidentiality agreement and the flow-down clauses are in place. Use this to answer "who is working on this engagement who is not ours".

      "Which subcontractors on Northwind have no executed contract?"

    • atlas_delivery_engagements_change_requests_createdelivery:write

      Raise a change request against an engagement. It is created as a draft with no impact assessed and no approvers, which is deliberate: assessing the impact and deciding the change belong to people. Use this to capture that a change has been requested, not to approve one.

      "Raise a change request for the extra migration run."

    • atlas_delivery_engagements_raid_items_createdelivery:write

      Log a risk, assumption, issue, dependency or decision on an engagement register. An issue has already happened, so its probability is always recorded as certain regardless of what is passed.

      "Log a risk that the data migration window may not hold."

    • atlas_delivery_engagements_createdelivery:write

      Create an engagement for a client account. It starts in the pipeline state; apply a playbook afterwards to set up its phases, deliverables and registers. Returns the new engagement with its id and reference. Use this when a person asks to open or set up a new engagement.

      "Open a new advisory engagement for the Northwind account."

    • atlas_delivery_engagements_updatedelivery:write

      Change the details of an engagement: its name, summary, type, fee model, values, dates, leadership and confidentiality. Send only the fields that change. Returns the updated engagement. The lifecycle state and the traffic lights have their own tools.

      "Move the end date of the Northwind engagement to the end of March."

    • atlas_delivery_engagements_deletedelivery:write

      Delete an engagement. It leaves every list and its code is released for reuse; the record is kept only for audit. Requires the administrator role. Archive it instead when it should only be put away, and use this only when the person explicitly asks to delete it.

      "Delete the duplicate Northwind engagement I created by mistake."

    • atlas_delivery_engagements_transitiondelivery:write

      Move an engagement to another lifecycle state, for example from mobilising to active or from active to on hold. It cannot close, terminate or cancel an engagement: ending an engagement belongs to a person, so those states are refused. The API also refuses a move the lifecycle does not allow and says why. Returns the engagement in its new state.

      "Put the Northwind engagement on hold while the client reorganises."

    • atlas_delivery_engagements_rag_setdelivery:write

      Set the traffic lights of an engagement: overall, schedule, budget, scope, quality, resourcing and client. Send only the lights that change, with the reason, which is recorded on the timeline. Returns the engagement with its new ratings.

      "Set the Northwind schedule light to amber because the data load slipped."

    • atlas_delivery_engagements_archivedelivery:write

      Archive an engagement. It leaves the default lists but keeps every record, and it can be restored. Returns the archived engagement.

      "Archive the finished Contoso pilot engagement."

    • atlas_delivery_engagements_restoredelivery:write

      Restore an archived engagement to the default lists. Returns the restored engagement.

      "Bring the archived Contoso pilot back."

    • atlas_delivery_engagements_templates_applydelivery:write

      Copy a playbook into an engagement: its phases, workstreams, deliverables, responsibility matrix, starter risks, meetings, information requests, closure items, report sections and checks. It writes across nine registers at once and cannot be undone, so confirm the playbook with the person first. Returns what was created.

      "Apply the due diligence playbook to the new Fabrikam engagement."

    • atlas_delivery_engagements_summary_getdelivery:read

      Count the engagements in the workspace: the total, the count in each lifecycle state, and how many are at risk. Use this to answer "how many engagements are active" without listing them.

      "How many engagements are active, and how many are at risk?"

    • atlas_delivery_my_engagements_listdelivery:read

      List the engagements the caller leads, as engagement partner, engagement manager, quality reviewer or PMO lead, most recently updated first. Use this to answer "which engagements am I responsible for".

      "Which engagements am I responsible for?"

    • atlas_delivery_portfolio_metrics_getdelivery:read

      The firm-wide delivery metrics across every engagement the caller can see: health, schedule and delivery performance in one view. Use this for a portfolio review rather than a question about one engagement.

      "How is the delivery portfolio performing overall?"

    • atlas_delivery_rollup_portfolio_getdelivery:read

      The portfolio rollup: the engagement counters and the most recently updated engagements the caller can see, in one read. Use this for a portfolio dashboard; use atlas_delivery_engagements_list to page through every engagement.

      "Give me the delivery portfolio dashboard."

    • atlas_delivery_rollup_raid_items_listdelivery:read

      The open risks, assumptions, issues and dependencies across every engagement the caller can see, with their scores. Use this to answer "what are the biggest risks across the portfolio".

      "What are the hottest open risks across all engagements?"

    • atlas_delivery_rollup_meetings_listdelivery:read

      The most recent governed meetings across every engagement the caller can see, with their dates and the state of their minutes. Use this to answer "whose minutes are still outstanding".

      "Which recent governance meetings still have minutes outstanding?"

    • atlas_delivery_rollup_info_requests_listdelivery:read

      The open information requests across every engagement the caller can see, including what is overdue. Use this to answer "what are we waiting on from clients".

      "What are we still waiting on from clients across the portfolio?"

    • atlas_delivery_rollup_actions_listdelivery:read

      The open actions across every engagement the caller can see, with owners and due dates. Use this to answer "what actions are outstanding".

      "Which engagement actions are still open across the firm?"

    • atlas_delivery_rollup_approvals_listdelivery:read

      The change request approval steps waiting for a decision across every engagement the caller can see. Use this to answer "which changes are waiting for approval".

      "Which change requests are waiting for approval?"

    • atlas_delivery_rollup_exports_listdelivery:read

      The document exports across every engagement the caller can see, with their state. Use this to find an export that was queued earlier.

      "Where is the status pack export I queued earlier?"

    • atlas_delivery_engagements_timeline_listdelivery:read

      Read the timeline of one engagement: every recorded event, newest first, with its kind and source. Use this to answer "what happened on this engagement" or "what changed this week".

      "What happened on the Northwind engagement this week?"

    • atlas_delivery_engagements_health_history_getdelivery:read

      Read how the health score of one engagement has moved over time. The live score is overwritten every hour, so use this to answer "has anything moved" rather than "how is it today".

      "Has the Northwind health score moved over the last month?"

    • atlas_delivery_engagements_custom_fields_getdelivery:read

      Read the custom field values of one engagement, with the field definitions the workspace uses. Use this before writing values, to learn which keys exist.

      "What custom fields are set on the Northwind engagement?"

    • atlas_delivery_engagements_custom_fields_setdelivery:write

      Write custom field values on one engagement. Only the keys sent are written, and a null clears a value. Returns every value after the write.

      "Set the Northwind sector field to Retail."

    • atlas_delivery_templates_listdelivery:read

      List the engagement playbooks available in the workspace: the built-in ones and any the workspace wrote or customised. Use this to choose a playbook before applying it to an engagement.

      "Which engagement playbooks do we have?"

    • atlas_delivery_templates_getdelivery:read

      Read one playbook in full: its phases, workstreams, deliverables, responsibility matrix, starter risks, meetings, information requests, closure items, report sections and checks.

      "Show me what the retainer playbook sets up."

    • atlas_delivery_templates_createdelivery:write

      Write a new playbook for the workspace, starting empty. Fill its collections afterwards with the collection tool. Requires the administrator role in client delivery.

      "Create a new playbook for forensic reviews."

    • atlas_delivery_templates_updatedelivery:write

      Change the name, headline, detail or typical duration of a playbook the workspace wrote or customised. Engagements already started from it are unchanged.

      "Change the typical duration of our forensic review playbook to twelve weeks."

    • atlas_delivery_templates_deletedelivery:write

      Delete a playbook the workspace wrote. A customised built-in is reset rather than deleted. Engagements already started from it are unchanged, because an engagement holds a copy.

      "Delete the forensic review playbook we no longer use."

    • atlas_delivery_templates_forkdelivery:write

      Take a copy of a built-in playbook so the workspace can edit it. Every engagement started from it afterwards takes the edited copy. Returns the workspace copy.

      "Make an editable copy of the built-in retainer playbook."

    • atlas_delivery_templates_resetdelivery:write

      Drop the workspace copy of a built-in playbook, restoring the built-in exactly. The edits in the copy are lost.

      "Reset our retainer playbook to the built-in version."

    • atlas_delivery_templates_collections_setdelivery:write

      Replace one collection of a playbook, for example its phases or its deliverables, with the list sent. The list replaces the collection whole, so an item left out is removed. Returns the playbook, or the rules the list breaks.

      "Replace the phases of our forensic review playbook with these four."

    • atlas_delivery_engagements_phases_listdelivery:read

      List the phases of one engagement in order, with their planned dates and gate state. Use this to answer "which phase are we in".

      "Which phase is the Northwind engagement in?"

    • atlas_delivery_engagements_phases_createdelivery:write

      Add a phase to an engagement, optionally ending at a gate that must be passed before the next phase starts. Returns the new phase.

      "Add a design phase to the Northwind engagement ending at a gate."

    • atlas_delivery_engagements_workstreams_createdelivery:write

      Add a workstream to an engagement, with its lead. Returns the new workstream.

      "Add a data migration workstream led by Priya."

    • atlas_delivery_engagements_team_members_listdelivery:read

      List the team of one engagement on every side: the firm, the client and third parties, with roles, allocation and key personnel. Use this to answer "who is on this engagement".

      "Who is on the Northwind engagement team?"

    • atlas_delivery_engagements_team_members_adddelivery:write

      Add a person to the team of an engagement, as a member, a client contact or an external party, with their role and allocation. Returns the new team entry.

      "Add Sam to the Northwind team as a consultant at fifty percent."

    • atlas_delivery_engagements_objectives_listdelivery:read

      List the objectives of one engagement with their success measures, baselines and targets. Use this to answer "what is this engagement meant to achieve".

      "What is the Northwind engagement meant to achieve?"

    • atlas_delivery_engagements_objectives_createdelivery:write

      Add an objective to an engagement with the measure of its success. Returns the new objective.

      "Add an objective to cut order processing time by thirty percent."

    • atlas_delivery_engagements_scope_statements_listdelivery:read

      List the scope of one engagement: what is in scope and what is explicitly excluded. Use this to answer "is this in scope".

      "Is warehouse integration in scope on Northwind?"

    • atlas_delivery_engagements_scope_statements_createdelivery:write

      Add a scope statement to an engagement, either something in scope or an explicit exclusion. Returns the new statement.

      "Record that payroll integration is out of scope."

    • atlas_delivery_engagements_stakeholders_setdelivery:write

      Add a stakeholder to an engagement, or update the one already recorded for the same person: their power, interest, sentiment and how they are managed. Returns the stakeholder.

      "Add the Northwind finance director as a sceptical, high-power stakeholder."

    • atlas_delivery_engagements_stakeholder_interactions_listdelivery:read

      List the recorded contacts with one stakeholder, newest first. Use this to answer "when did we last speak to them".

      "When did we last speak to the Northwind finance director?"

    • atlas_delivery_engagements_stakeholder_interactions_logdelivery:write

      Record a contact with a stakeholder: the channel, subject, summary and how they felt afterwards. Returns the recorded contact.

      "Log today's call with the Northwind sponsor: she is supportive of the new plan."

    • atlas_delivery_engagements_comments_listcomments:read

      List the comments on one engagement, oldest first, one page at a time. Use this to read the discussion on an engagement.

      "Show me the discussion on the Northwind engagement."

    • atlas_delivery_engagements_comments_createcomments:write

      Post a comment on an engagement, or a reply to another comment. Returns the new comment.

      "Comment on the Northwind engagement that the kickoff moved to Tuesday."

    • atlas_delivery_engagements_attachments_listattachments:read

      List the files attached to one engagement, with their names, sizes and who can see them.

      "What files are attached to the Northwind engagement?"

    • atlas_delivery_engagements_attachments_uploadattachments:write

      Start uploading a file to an engagement. Returns an upload URL and an attachment id: send the file bytes to the URL, then call atlas_delivery_engagements_attachments_finalize with the id.

      "Attach this statement of work to the Northwind engagement."

    • atlas_delivery_engagements_attachments_finalizeattachments:write

      Finish an upload started with atlas_delivery_engagements_attachments_upload, once the bytes have been sent. Returns the attachment.

      "Finish attaching the statement of work I just uploaded."

    • atlas_delivery_engagements_attachments_downloadattachments:read

      Get a short-lived download URL for a file attached to an engagement. Returns the URL and when it expires.

      "Give me a download link for the Northwind statement of work."

    • atlas_delivery_engagements_attachments_deleteattachments:delete

      Delete a file attached to an engagement, removing it from the engagement file list. Use this only when the person asks to delete it.

      "Remove the outdated org chart from the Northwind files."

    • atlas_delivery_engagements_attachments_share_revokedelivery:write

      Take an engagement file back from the client portal, so only the firm sees it. The change is audited and recorded on the timeline.

      "Take the draft org chart back from the Northwind client portal."

    • atlas_delivery_engagements_meeting_series_listdelivery:read

      List the recurring governed meetings of one engagement, such as a steering committee or a weekly check-in, with their chair, quorum rules and minutes rules.

      "What recurring governance meetings does Northwind have?"

    • atlas_delivery_engagements_meeting_series_createdelivery:write

      Set up a recurring governed meeting on an engagement, with its chair, secretary, quorum policy and minutes rules. Returns the new series.

      "Set up a monthly steering committee for Northwind."

    • atlas_delivery_engagements_meeting_series_updatedelivery:write

      Change a recurring governed meeting: its name, people, quorum policy, minutes rules or whether it is still active. Send only the fields that change. Returns the updated series.

      "Make Priya the chair of the Northwind steering committee."

    • atlas_delivery_engagements_meeting_series_deletedelivery:write

      Delete a recurring governed meeting that has no meetings. A series that has meetings is refused; set it inactive with the update tool instead, which keeps its history.

      "Delete the steering committee series I set up by mistake."

    • atlas_delivery_engagements_meeting_series_previewdelivery:read

      Preview the next occurrences of a recurring governed meeting without scheduling anything. Use this to check the dates before creating meetings from the series.

      "When would the next six steering committees fall?"

    • atlas_delivery_engagements_meeting_series_terms_getdelivery:read

      Read the terms of reference of a recurring governed meeting: its purpose, membership and decision rights.

      "What are the terms of reference of the Northwind steering committee?"

    • atlas_delivery_engagements_meeting_series_terms_setdelivery:write

      Replace the terms of reference of a recurring governed meeting. Returns the saved terms.

      "Update the steering committee terms of reference with this text."

    • atlas_delivery_engagements_meetings_calendar_getdelivery:read

      Read the governed meetings of one engagement between two dates, for a calendar view. Use this to answer "what meetings are there next week".

      "What Northwind meetings are there next week?"

    • atlas_delivery_engagements_meetings_createdelivery:write

      Schedule a governed meeting on an engagement, on its own or as part of a series. Returns the new meeting.

      "Schedule a risk review for Northwind on Thursday at ten."

    • atlas_delivery_engagements_series_meetings_createdelivery:write

      Schedule the next meeting of a recurring series, taking its people, quorum rules and minutes rules from the series. Returns the new meeting.

      "Schedule the next Northwind steering committee."

    • atlas_delivery_engagements_meetings_getdelivery:read

      Read one governed meeting: its date, chair, attendees, quorum and the state of its minutes. Quorum has three states: met, not met, and not evaluated.

      "Show me the details of the last Northwind steering committee."

    • atlas_delivery_engagements_meetings_updatedelivery:write

      Change a governed meeting: its title, time, length, people, location or quorum. Send only the fields that change. Returns the updated meeting.

      "Move the Northwind risk review to Friday afternoon."

    • atlas_delivery_engagements_meetings_deletedelivery:write

      Take a governed meeting off the register. It is refused once the minutes have been circulated; cancel the meeting instead, which keeps the record and says it did not go ahead.

      "Delete the duplicate risk review I scheduled."

    • atlas_delivery_engagements_meetings_linkdelivery:write

      Link a governed meeting to an Atlas meeting, so its notes and recording sit with the governed record. Returns the linked meeting.

      "Link the steering committee to the Atlas meeting we recorded."

    • atlas_delivery_engagements_meetings_unlinkdelivery:write

      Remove the link between a governed meeting and an Atlas meeting. Neither meeting is deleted.

      "Unlink the steering committee from the wrong Atlas meeting."

    • atlas_delivery_engagements_meeting_minutes_importdelivery:write

      Import the notes of the linked meeting from Granola into the minutes of a governed meeting. The workspace must have Granola connected. Returns the minutes.

      "Pull the Granola notes into the steering committee minutes."

    • atlas_delivery_engagements_meetings_startdelivery:write

      Mark a governed meeting as in progress. Returns the meeting.

      "Mark the steering committee as started."

    • atlas_delivery_engagements_meetings_held_markdelivery:write

      Record that a governed meeting took place, with when and for how long. Whether it is recorded as held or as held without quorum follows from the quorum already recorded, not from the caller. Returns the meeting.

      "Record that the steering committee took place this morning for ninety minutes."

    • atlas_delivery_engagements_meetings_canceldelivery:write

      Cancel a governed meeting. The record stays, marked cancelled, with the reason.

      "Cancel Thursday's risk review because the sponsor is away."

    • atlas_delivery_engagements_meeting_carried_items_listdelivery:read

      List the agenda items and actions carried forward into this meeting from earlier meetings of the same series. Use this to answer "what is still open from last time".

      "What is still open from the last steering committee?"

    • atlas_delivery_engagements_meeting_agenda_items_listdelivery:read

      Read the agenda of a governed meeting in order, with each item's purpose, timebox, status and outcome.

      "What is on the agenda for the steering committee?"

    • atlas_delivery_engagements_meeting_agenda_items_createdelivery:write

      Add an item to the agenda of a governed meeting, optionally tied to a RAID item or a deliverable. Returns the new item.

      "Add the budget reforecast to the steering committee agenda for twenty minutes."

    • atlas_delivery_engagements_meeting_agenda_items_reorderdelivery:write

      Put the agenda of a governed meeting in a new order. Send every item id in the order wanted. Returns the agenda.

      "Put the budget item first on the steering committee agenda."

    • atlas_delivery_engagements_meeting_agenda_items_updatedelivery:write

      Change an agenda item: its title, purpose, timebox, time taken, status or outcome. Send only the fields that change. Returns the item.

      "Mark the budget agenda item as covered."

    • atlas_delivery_engagements_meeting_agenda_items_deletedelivery:write

      Remove an item from the agenda of a governed meeting. Use defer instead when the item should move to the next meeting.

      "Take the vendor update off the agenda."

    • atlas_delivery_engagements_meeting_agenda_items_deferred_markdelivery:write

      Defer an agenda item, so it is carried forward to the next meeting of the series. Returns the item.

      "Defer the vendor update to the next steering committee."

    • atlas_delivery_engagements_meeting_raid_items_logdelivery:write

      Log a risk, assumption, issue or dependency raised under an agenda item, linked back to the meeting. Returns the new RAID item.

      "Log a risk from the budget discussion: the reforecast may exceed the cap."

    • atlas_delivery_engagements_meeting_decisions_logdelivery:write

      Log a decision taken under an agenda item in the decision register, linked back to the meeting. Returns the new decision.

      "Log the decision to extend testing by two weeks."

    • atlas_delivery_engagements_meeting_agenda_items_convertdelivery:write

      Turn an agenda item into a task, linked back to the meeting. Returns the new task.

      "Turn the vendor follow-up agenda item into a task for Sam."

    • atlas_delivery_engagements_meeting_agenda_outcomes_recorddelivery:write

      Record what was concluded on an agenda item, and whether a decision was made. Returns the item.

      "Record that the committee agreed the revised timeline."

    • atlas_delivery_engagements_meeting_attendees_listdelivery:read

      List the attendees of a governed meeting with their roles, whether they vote, and their attendance.

      "Who is invited to the steering committee?"

    • atlas_delivery_engagements_meeting_attendees_adddelivery:write

      Invite someone to a governed meeting, as a member, a client contact or an external party, with their role and whether they vote. Returns the attendee.

      "Invite the Northwind CFO to the steering committee as a voting member."

    • atlas_delivery_engagements_meeting_attendees_updatedelivery:write

      Change an attendee's role, or whether they vote or are required. Returns the attendee.

      "Make the CFO a non-voting observer."

    • atlas_delivery_engagements_meeting_attendees_removedelivery:write

      Remove an attendee from a governed meeting, for example someone invited by mistake. Use the attendance tool to record apologies instead.

      "Remove the guest I invited to the steering committee by mistake."

    • atlas_delivery_engagements_meeting_attendee_delegates_assigndelivery:write

      Record that an attendee sends someone else in their place. Returns the delegate.

      "The CFO is sending her deputy to the steering committee instead."

    • atlas_delivery_engagements_meeting_attendance_recorddelivery:write

      Record an attendee's response or attendance, for example accepted, attended or sent apologies. Returns the attendee.

      "Record that the sponsor sent apologies."

    • atlas_delivery_engagements_meeting_votes_listdelivery:read

      List the votes cast in a governed meeting, by motion, with who voted and how. Use this to answer "was the motion carried".

      "Was the motion to extend testing carried?"

    • atlas_delivery_engagements_meeting_votes_recorddelivery:write

      Record a vote cast by a voting attendee on a motion. Returns the vote.

      "Record the CFO vote in favour of extending testing."

    • atlas_delivery_engagements_meeting_minutes_getdelivery:read

      Read the minutes of a governed meeting with their approval state.

      "Show me the minutes of the last steering committee."

    • atlas_delivery_engagements_meeting_minutes_draftdelivery:write

      Write the first draft of the minutes of a governed meeting, and optionally circulate them for approval straight away. Returns the minutes.

      "Draft the steering committee minutes from these notes."

    • atlas_delivery_engagements_meeting_minutes_updatedelivery:write

      Save a new version of the minutes of a governed meeting. Each save is kept as a version. Returns the minutes.

      "Save this revised version of the minutes."

    • atlas_delivery_engagements_meeting_minutes_versions_listdelivery:read

      List the saved versions of the minutes of a governed meeting, newest first.

      "How many times have the steering committee minutes been revised?"

    • atlas_delivery_engagements_meeting_minutes_comparedelivery:read

      Compare the latest minutes of a governed meeting with the version before, and return what changed.

      "What changed in the latest version of the minutes?"

    • atlas_delivery_engagements_plan_items_createdelivery:write

      Add a row to the plan of an engagement: a task, a milestone, a summary or a gate, with its dates, effort and constraints. Returns the new plan item.

      "Add a user acceptance testing task to the Northwind plan for three weeks."

    • atlas_delivery_engagements_plan_items_updatedelivery:write

      Change a plan item: its title, dates, progress, effort, parent or constraints. Send only the fields that change. Returns the item.

      "Mark the data load task as sixty percent complete."

    • atlas_delivery_engagements_plan_items_deletedelivery:write

      Delete a plan item and the dependencies that touch it. Returns how many dependencies were removed. Use this only when the person asks to delete it.

      "Delete the duplicate testing task from the plan."

    • atlas_delivery_engagements_plan_items_convertdelivery:write

      Create a task from a plan item and link the two, so progress on the task shows in the plan. Returns the new task.

      "Turn the data load plan item into a task."

    • atlas_delivery_engagements_plan_items_reorderdelivery:write

      Move plan items to new positions or under new parents in one request. Returns the plan.

      "Move the testing tasks under the build phase."

    • atlas_delivery_engagements_plan_items_movedelivery:write

      Shift the dates of several plan items by the same number of days. Returns the moved items.

      "Push every testing task back five days."

    • atlas_delivery_engagements_plan_items_importdelivery:write

      Bring existing tasks into the plan of an engagement as plan items linked to them. Returns the new plan items.

      "Bring these three project tasks into the Northwind plan."

    • atlas_delivery_engagements_plan_dependencies_listdelivery:read

      List the dependencies between plan items of one engagement, with their type and lag.

      "What depends on the data load task?"

    • atlas_delivery_engagements_plan_dependencies_createdelivery:write

      Add a dependency between two plan items. The API refuses a dependency its scheduling rules do not allow and says why. Returns the new dependency.

      "Make testing start only after the data load finishes."

    • atlas_delivery_engagements_plan_dependencies_validatedelivery:write

      Check whether a dependency could be added, without adding it. Returns the refusal the create would give, or nothing when it would succeed.

      "Could I make go-live depend on training without breaking the plan?"

    • atlas_delivery_engagements_plan_dependencies_updatedelivery:write

      Change the type, lag or hardness of a dependency. Returns the dependency.

      "Add a two-day lag to the testing dependency."

    • atlas_delivery_engagements_plan_dependencies_deletedelivery:write

      Remove a dependency between two plan items. Recompute the critical path afterwards to see the effect.

      "Remove the dependency between training and go-live."

    • atlas_delivery_engagements_plan_gantt_getdelivery:read

      Read the plan, its dependencies and its variance against the baseline in one response, as a Gantt chart needs them.

      "Show me the Northwind plan as a Gantt chart."

    • atlas_delivery_engagements_critical_path_getdelivery:read

      Read the critical path of the plan in the order it runs, as last computed. It does not recompute; use the refresh tool for that.

      "What is on the Northwind critical path?"

    • atlas_delivery_engagements_critical_path_refreshdelivery:write

      Recompute the critical path and the float of every plan item. Returns the new critical path.

      "Recompute the Northwind critical path."

    • atlas_delivery_engagements_plan_float_listdelivery:read

      List plan items by how much float they have, least first. Use this to answer "what is about to become critical".

      "Which tasks are about to become critical?"

    • atlas_delivery_engagements_plan_exportdelivery:read

      Export the plan of an engagement with every item, dependency and date, in a form another scheduling tool can import.

      "Export the Northwind plan so I can open it in another scheduling tool."

    • atlas_delivery_engagements_plan_rescheduledelivery:write

      Reschedule the plan from its dependencies and constraints. By default it returns the proposed moves and changes nothing; send apply with the fingerprint the proposal reported to write them.

      "What would change if we rescheduled the plan from next Monday?"

    • atlas_delivery_engagements_plan_leveling_rundelivery:write

      Level the plan so no person is booked beyond capacity. By default it returns the proposed moves and changes nothing; send apply with the fingerprint the proposal reported to write them.

      "Is anyone overbooked on Northwind, and what would levelling move?"

    • atlas_delivery_engagements_milestone_trend_getdelivery:read

      Read the milestone trend: how the forecast date of each milestone has moved at each capture. Use this to answer "are the milestones slipping".

      "Are the Northwind milestones slipping?"

    • atlas_delivery_engagements_milestone_trend_capturedelivery:write

      Capture today's forecast date of every milestone as a point on the milestone trend. Returns the capture.

      "Capture this week's milestone forecast."

    • atlas_delivery_engagements_baselines_listdelivery:read

      List the baselines of one engagement, marking which is current.

      "Which baselines does Northwind have?"

    • atlas_delivery_engagements_baselines_capturedelivery:write

      Capture the plan as it stands as a named baseline: the original, a rebaseline after a change, or a what-if. Returns the baseline.

      "Capture a what-if baseline before we replan."

    • atlas_delivery_engagements_baselines_getdelivery:read

      Read one baseline with the dates of every plan item it captured.

      "Show me the original Northwind baseline."

    • atlas_delivery_engagements_baselines_deletedelivery:write

      Delete a baseline that is no longer needed, such as a what-if. The current baseline cannot be deleted.

      "Delete the what-if baseline we no longer need."

    • atlas_delivery_engagements_plan_variance_getdelivery:read

      Compare the plan with the current baseline: how far each item and the whole plan has moved. Use this to answer "how far behind are we".

      "How far behind the baseline is Northwind?"

    • atlas_delivery_engagements_baseline_variance_getdelivery:read

      Compare the plan with one named baseline rather than the current one. Use this to answer "what did the last replan cost".

      "What did the last replan cost against the original baseline?"

    • atlas_delivery_engagements_change_requests_summary_getdelivery:read

      Summarise the change register of one engagement: the count in each status, and the value sitting in the approval chain that has not landed yet.

      "How much change value is waiting in the approval chain on Northwind?"

    • atlas_delivery_engagements_change_requests_getdelivery:read

      Read one change request with its impact assessment, approval steps and history.

      "Show me the change request for the second workshop."

    • atlas_delivery_engagements_change_requests_updatedelivery:write

      Change the title, description, type, urgency or flags of a change request. Send only the fields that change. Returns the change request.

      "Make the second workshop change request high urgency."

    • atlas_delivery_engagements_change_requests_deletedelivery:write

      Delete a change request that is still open. A decided change is part of the record and is refused; withdraw an open change instead when its history should stay.

      "Delete the change request I raised twice."

    • atlas_delivery_engagements_change_requests_submitdelivery:write

      Submit a change request for review. Returns the change request in its new state.

      "Submit the second workshop change request for review."

    • atlas_delivery_engagements_change_requests_reviewdelivery:write

      Start the review of a submitted change request, so its approvers can decide. Returns the change request in its new state.

      "Start the review of the second workshop change request."

    • atlas_delivery_engagements_change_requests_withdrawdelivery:write

      Withdraw a change request, with the reason. The record stays. Returns the change request.

      "Withdraw the second workshop change request; the client dropped it."

    • atlas_delivery_engagements_change_requests_deferred_markdelivery:write

      Defer a change request to be considered later, with the reason. Returns the change request.

      "Defer the reporting change request until phase two."

    • atlas_delivery_engagements_change_requests_share_revokedelivery:write

      Take a change request back from the client portal, with the reason. The change is audited and recorded on the timeline.

      "Take the pricing change request back from the client portal."

    • atlas_delivery_engagements_change_requests_completedelivery:write

      Mark an approved change request as implemented. Returns the change request.

      "Mark the approved workshop change as implemented."

    • atlas_delivery_engagements_change_approvals_listdelivery:read

      List the approval steps of a change request with each approver and their decision.

      "Who still has to decide on the second workshop change?"

    • atlas_delivery_engagements_change_approvers_adddelivery:write

      Add an approver to a change request at a given step, as a member, a client contact or an external party. Returns the approval step.

      "Add the Northwind sponsor as an approver on the workshop change."

    • atlas_delivery_engagements_status_reports_getdelivery:read

      Read one status report in full: its period, traffic lights, narrative, sections and state.

      "Show me last week's Northwind status report."

    • atlas_delivery_engagements_status_reports_createdelivery:write

      Start a blank status report for a period. Returns the draft.

      "Start a blank status report for this week."

    • atlas_delivery_engagements_status_reports_generatedelivery:write

      Start a status report for a period, pre-filled from the engagement: traffic lights, progress, RAID highlights and milestones. Returns the draft for a person to edit.

      "Draft this week's Northwind status report from the engagement."

    • atlas_delivery_engagements_status_reports_updatedelivery:write

      Edit a draft status report: its traffic lights, summary, accomplishments, next period, asks of the client and decisions needed. Send only the fields that change. Returns the report.

      "Set the executive summary of the draft report to this paragraph."

    • atlas_delivery_engagements_status_reports_deletedelivery:write

      Delete a status report that has not been issued. An issued report is refused, because the client holds it; revise it instead.

      "Delete the draft status report I started by mistake."

    • atlas_delivery_engagements_status_reports_since_last_getdelivery:read

      What has changed on the engagement since the last issued status report: traffic light moves, new risks, completed work and slipped dates. Use this when writing the next report.

      "What has changed since the last status report?"

    • atlas_delivery_engagements_status_report_sections_listdelivery:read

      List the sections of a status report in order, with each section's traffic light, progress, owner and text.

      "What sections are in this week's status report?"

    • atlas_delivery_engagements_status_report_sections_createdelivery:write

      Add a section to a draft status report, for example a workstream update or a budget section. Returns the section.

      "Add a data migration workstream section to the draft report."

    • atlas_delivery_engagements_status_report_sections_updatedelivery:write

      Edit a section of a draft status report. Send only the fields that change. Returns the section.

      "Change the data migration section light to amber."

    • atlas_delivery_engagements_status_report_sections_deletedelivery:write

      Remove a section from a draft status report. A section of an issued report cannot be removed.

      "Remove the appendix section from the draft report."

    • atlas_delivery_engagements_status_reports_refreshdelivery:write

      Pull the latest figures onto a draft status report. Only the figures are refreshed; every word the author wrote is kept. Returns the report.

      "Refresh the figures on the draft status report."

    • atlas_delivery_engagements_status_reports_submitdelivery:write

      Submit a draft status report for approval. Returns the report in its new state.

      "Submit the draft status report for approval."

    • atlas_delivery_engagements_status_reports_share_revokedelivery:write

      Take a status report back from the client, with the reason. The change is audited and recorded on the timeline.

      "Take last week's status report back from the client."

    • atlas_delivery_engagements_status_report_revisions_createdelivery:write

      Revise an issued status report: the issued one is marked superseded and a new draft for the same period is created from the engagement as it stands. Returns the new draft.

      "Start a corrected version of last week's issued report."

    • atlas_delivery_engagements_status_report_recipients_listdelivery:read

      List who a status report is sent to, with the channel and format for each.

      "Who receives the Northwind status report?"

    • atlas_delivery_engagements_status_report_recipients_adddelivery:write

      Add a recipient to a status report, as a member, a client contact or an external party, with the channel and format. Returns the recipient.

      "Add the Northwind CFO to the status report recipients by email."

    • atlas_delivery_engagements_deliverables_createdelivery:write

      Add a deliverable to an engagement, with its due date, owner, review tier and acceptance criteria. Returns the new deliverable.

      "Add the target operating model as a contractual deliverable due in May."

    • atlas_delivery_engagements_deliverables_getdelivery:read

      Read one deliverable with its stage, versions, reviews, criteria and sign-off.

      "Show me the target operating model deliverable."

    • atlas_delivery_engagements_acceptance_blockers_listdelivery:read

      List everything stopping a deliverable from being accepted: unmet criteria, open review notes, missing reviews and missing sign-off. An empty list means it can be accepted.

      "What is stopping the operating model deliverable from being accepted?"

    • atlas_delivery_engagements_deliverable_certificates_getdelivery:read

      Return the API path of the acceptance certificate for a deliverable, in the format asked for. Compose it with the API base URL and an Authorization header to download the document.

      "Get the acceptance certificate for the operating model deliverable as a PDF."

    • atlas_delivery_engagements_deliverable_versions_capturedelivery:write

      Record a new version of a deliverable, with what changed and the file it lives in. Returns the version.

      "Record version two of the operating model with this change log."

    • atlas_delivery_engagements_deliverable_criteria_listdelivery:read

      List the acceptance criteria of a deliverable with whether each is met, not met or waived.

      "Which acceptance criteria on the operating model are still unmet?"

    • atlas_delivery_engagements_deliverable_criteria_createdelivery:write

      Add an acceptance criterion to a deliverable. Returns the criterion.

      "Add an acceptance criterion that every process has an owner."

    • atlas_delivery_engagements_deliverable_reviews_listdelivery:read

      List the reviews of a deliverable, one per gate, with their reviewer and outcome.

      "Which reviews has the operating model been through?"

    • atlas_delivery_engagements_deliverable_reviews_requestdelivery:write

      Request a review of a deliverable at a gate, for example quality review or legal review. Returns the review.

      "Request a quality review of the operating model from Priya."

    • atlas_delivery_engagements_review_checklists_getdelivery:read

      Read the checklist of a deliverable review with the answer to each question.

      "Show me the quality review checklist for the operating model."

    • atlas_delivery_engagements_review_checklists_setdelivery:write

      Replace the questions of a deliverable review checklist with the list sent. Returns the checklist.

      "Use these ten questions as the review checklist."

    • atlas_delivery_engagements_sign_offs_requestdelivery:write

      Ask the client signatory to sign off a deliverable, by the method chosen, with a due date and an optional lapse into deemed acceptance. Returns the sign-off request.

      "Ask the Northwind sponsor to sign off the operating model by Friday."

    • atlas_delivery_engagements_commercials_effort_getdelivery:read

      Read the effort of one engagement: actual, billable and non-billable effort in minutes, utilisation and leverage.

      "How much of the effort on Northwind has been billable?"

    • atlas_delivery_engagements_rate_cards_listdelivery:read

      List the rate cards of one engagement with their currency and dates.

      "Which rate cards does Northwind use?"

    • atlas_delivery_engagements_rate_cards_createdelivery:write

      Add a rate card to an engagement, with its currency, geography and the dates it applies. Add its rates afterwards. Returns the rate card.

      "Add a UK rate card in pounds from January."

    • atlas_delivery_engagements_rate_card_entries_listdelivery:read

      List the bill rates on a rate card by grade and role. The internal cost rates are never returned by this tool, because they are the firm's own commercial position.

      "What are the bill rates on the UK rate card?"

    • atlas_delivery_engagements_expenses_listdelivery:read

      List the expenses of one engagement, with filters for status, billable and policy breaches.

      "Which Northwind expenses breach the policy?"

    • atlas_delivery_engagements_expenses_createdelivery:write

      Record an expense against an engagement, as a draft. Returns the expense.

      "Record a 240 pound train fare to the client site yesterday."

    • atlas_delivery_engagements_expenses_updatedelivery:write

      Change a draft expense. Send only the fields that change. Returns the expense.

      "Mark the train fare as billable."

    • atlas_delivery_engagements_expenses_submitdelivery:write

      Submit a draft expense for approval. A submitted expense cannot be edited until it is rejected back. Returns the expense.

      "Submit my train fare expense for approval."

    • atlas_delivery_engagements_expenses_disbursedelivery:write

      Record that an approved expense was reimbursed to the person who paid it. Returns the expense.

      "Record that the train fare was reimbursed."

    • atlas_delivery_engagements_billing_milestones_listdelivery:read

      List the billing milestones of one engagement with their trigger, amount and state: planned, earned, invoiced, paid, disputed or on hold.

      "Which Northwind billing milestones are not yet invoiced?"

    • atlas_delivery_engagements_billing_milestones_createdelivery:write

      Add a billing milestone to an engagement, with what makes it billable and the amount. Returns the milestone.

      "Add a 50,000 billing milestone triggered by acceptance of the operating model."

    • atlas_delivery_engagements_billing_milestones_invoiced_markdelivery:write

      Mark a billing milestone as invoiced, with the invoice reference. Returns the milestone.

      "Mark the design milestone as invoiced on INV-2291."

    • atlas_delivery_engagements_billing_milestones_paid_markdelivery:write

      Mark an invoiced billing milestone as paid. Returns the milestone.

      "Mark the design milestone as paid."

    • atlas_delivery_engagements_billing_milestones_disputed_markdelivery:write

      Mark a billing milestone as disputed by the client, with the reason. Returns the milestone.

      "The client disputes the design milestone invoice."

    • atlas_delivery_engagements_billing_milestones_held_markdelivery:write

      Put a billing milestone on hold, with the reason. Returns the milestone.

      "Put the build milestone on hold until the purchase order arrives."

    • atlas_delivery_engagements_revenue_events_listdelivery:read

      List the revenue recognition events of one engagement, newest first.

      "What revenue recognition events has Northwind had?"

    • atlas_delivery_engagements_acceptance_getdelivery:read

      Read the client acceptance record of one engagement: its risk tier, the facts it was assessed on, its checks and the decision.

      "Show me the client acceptance record for Northwind."

    • atlas_delivery_engagements_acceptance_createdelivery:write

      Open the client acceptance record for an engagement, with the facts the risk assessment rests on. The checks those facts require are created with it. Returns the record.

      "Open client acceptance for Northwind as a high-risk, cross-border engagement."

    • atlas_delivery_engagements_acceptance_updatedelivery:write

      Change the facts of the client acceptance record. The checks are brought in line with the new facts. Send only the fields that change. Returns the record.

      "Record that the Northwind work will use subcontractors."

    • atlas_delivery_engagements_acceptance_reviewdelivery:write

      Move the client acceptance record into review, so the decision can be taken. Returns the record.

      "Move the Northwind acceptance into review."

    • atlas_delivery_engagements_acceptance_checks_listdelivery:read

      List the acceptance checks of one engagement, such as conflicts, sanctions and anti-money-laundering checks, with their state and expiry.

      "Which acceptance checks are still outstanding on Northwind?"

    • atlas_delivery_engagements_acceptance_checks_createdelivery:write

      Add an acceptance check to an engagement, for example a sanctions screen on a named party. Returns the check.

      "Add a sanctions check on Northwind Holdings."

    • atlas_delivery_engagements_acceptance_check_terms_updatedelivery:write

      Change the terms a check runs under: the export control restrictions and the enhanced due diligence facts. Send only the fields that change. Returns the check.

      "Record that only US persons may work on the export control check."

    • atlas_delivery_engagements_closure_readiness_getdelivery:read

      Answer "can this engagement close". Returns whether it is ready, every blocker and warning at once, and the closure checklist it was judged against.

      "Can the Northwind engagement close yet?"

    • atlas_delivery_engagements_reopendelivery:write

      Reopen a closed engagement, with the reason. Requires the owner or administrator role. Returns the engagement.

      "Reopen the closed Contoso engagement for the warranty fix."

    • atlas_delivery_engagements_closure_items_listdelivery:read

      List the closure checklist of one engagement with the state of each item.

      "What is left on the Northwind closure checklist?"

    • atlas_delivery_engagements_closure_items_generatedelivery:write

      Create the standard closure checklist for an engagement. An engagement that already has one is left unchanged. Returns the checklist.

      "Create the closure checklist for Northwind."

    • atlas_delivery_engagements_closure_items_reopendelivery:write

      Reopen a closure checklist item that was done or waived. Returns the item.

      "Reopen the archive closure item; the files were not all moved."

    • atlas_delivery_engagements_approvals_listdelivery:read

      Read the approval history of one record with its current position: whether the version as it stands is approved. A record can have approvals against earlier versions and still not be approved as it stands.

      "Is the current version of the operating model approved?"

    • atlas_delivery_engagements_raid_items_getdelivery:read

      Read one RAID item in full: its kind, status, scores, heat band, owner, response plans, review dates and client sharing state. Use this after listing the register to answer a question about one risk, assumption, issue, dependency or decision.

      "Show me the full detail of the data migration risk."

    • atlas_delivery_engagements_raid_items_updatedelivery:write

      Change the fields of one RAID item, such as its owner, response, mitigation and contingency plans, review cadence, dates, cost and schedule impact, or decision rationale. Only the fields sent change; send null to clear a field. Returns the updated item.

      "Set the mitigation plan on the vendor risk and make Priya the owner."

    • atlas_delivery_engagements_raid_items_deletedelivery:write

      Delete a RAID item that was logged in error. An item that has a history worth keeping should be voided instead, with atlas_delivery_engagements_raid_items_void. Returns the result of the deletion.

      "Delete the duplicate risk I logged a minute ago."

    • atlas_delivery_engagements_raid_items_bulk_createdelivery:write

      Log up to fifty RAID items on one engagement in a single request. Either every item is created or none is. Returns the created items.

      "Log these five risks from the workshop on the engagement."

    • atlas_delivery_engagements_raid_items_bulk_reviewdelivery:write

      Record a review of several RAID items at once, optionally with a shared note and a new review cadence. Each item gets its next review date moved forward. Returns the reviewed items.

      "Mark all the risks we went through in steering as reviewed."

    • atlas_delivery_engagements_raid_items_ageing_getdelivery:read

      Report how long open RAID items have been open, oldest first, optionally for one kind. Use this to answer "which issues have been sitting there longest".

      "Which open issues are the oldest on this engagement?"

    • atlas_delivery_engagements_raid_items_due_reviews_listdelivery:read

      List RAID items whose review falls due within the given number of days, including those already overdue. Use this to prepare a risk review.

      "Which risks are due for review this week?"

    • atlas_delivery_engagements_raid_items_exportdelivery:read

      Render the RAID log as a Markdown document, for the internal team or in the client-safe form. Returns the Markdown text, or a job ticket when the log is too large to render in one request; poll a ticket with atlas_delivery_engagements_export_jobs_get.

      "Give me the client version of the RAID log."

    • atlas_delivery_engagements_raid_items_heatmap_getdelivery:read

      Read the probability by impact heat map of open risks and issues, with the count in each cell. Use this to answer "how exposed is this engagement".

      "Show me the risk heat map for this engagement."

    • atlas_delivery_engagements_raid_items_importdelivery:write

      Seed the RAID register from the risk log of the delivery project linked to the engagement. Items already imported are not duplicated. Returns what was imported.

      "Pull the risks from the linked project into the engagement register."

    • atlas_delivery_engagements_raid_items_summary_getdelivery:read

      Read the RAID register summary: counts by kind, status and heat band, overdue reviews, and escalated items. Use this for a quick picture before reading individual items.

      "Summarise the RAID register."

    • atlas_delivery_engagements_raid_items_convertdelivery:write

      Change the kind of a RAID item, for example when a risk has occurred and becomes an issue. The history is kept. Returns the converted item.

      "The supplier risk has happened, turn it into an issue."

    • atlas_delivery_engagements_raid_items_change_requests_createdelivery:write

      Raise a draft change request from a RAID item and link the two. The change request is created as a draft; assessing and approving it stays with people. Returns the change request.

      "Raise a change request from the scope creep issue."

    • atlas_delivery_engagements_raid_items_tasks_createdelivery:write

      Create an Atlas task from a RAID item, so the action to address it is tracked with the rest of the work. Returns the created task.

      "Create a task for Sam to chase the data access dependency by Friday."

    • atlas_delivery_engagements_raid_items_escalatedelivery:write

      Escalate a RAID item to a named person, an escalation level, or an escalation path, with the reason. The escalation is recorded on the item history. Returns the updated item.

      "Escalate the budget risk to the engagement partner."

    • atlas_delivery_engagements_raid_items_events_listdelivery:read

      Read the history of one RAID item: every status change, score change, review, escalation and share, with who did it and when.

      "What has happened to the vendor risk since it was logged?"

    • atlas_delivery_engagements_raid_items_reviewdelivery:write

      Record that a RAID item was reviewed, with an optional note and a new review cadence. The next review date moves forward. Returns the updated item.

      "Mark the migration risk as reviewed with a note that nothing changed."

    • atlas_delivery_engagements_raid_items_scoredelivery:write

      Rescore a RAID item: its probability and impact, and optionally the residual probability and impact after the response. The heat band is recalculated. Returns the updated item.

      "Raise the probability of the resourcing risk to 4."

    • atlas_delivery_engagements_raid_items_transitiondelivery:write

      Move a RAID item between its working states: open, in progress, mitigating, or monitoring. Closing states are set with atlas_delivery_engagements_raid_items_resolve. Returns the updated item.

      "Move the vendor risk to mitigating."

    • atlas_delivery_engagements_raid_items_share_revokedelivery:write

      Stop sharing a RAID item with the client, with the reason. The item disappears from client views and client exports. Returns the updated item.

      "Stop sharing the pricing issue with the client."

    • atlas_delivery_engagements_raid_items_voiddelivery:write

      Void a RAID item that should never have been raised, with the reason. A voided item stays in the history but leaves every count. Returns the updated item.

      "Void the duplicate issue, it was logged twice."

    • atlas_delivery_engagements_raid_items_resolvedelivery:write

      Close a RAID item as resolved or closed, with a resolution summary. The route also accepts, rejects, validates and invalidates items, but each of those is a person putting their name to a judgement, so this tool offers only RESOLVED and CLOSED. Returns the updated item.

      "Close the access issue as resolved, access was granted on Monday."

    • atlas_delivery_engagements_info_requests_createdelivery:write

      Add an information request to an engagement: what is asked of the client, in what format, by when, and whether it blocks the plan. It is created unsent; send it with atlas_delivery_engagements_info_requests_send. Returns the created request.

      "Ask the client for the last three years of management accounts by the 14th."

    • atlas_delivery_engagements_info_requests_bulk_createdelivery:write

      Add up to two hundred information requests to an engagement in one request, optionally all under one category. Either every request is created or none is. Returns the created requests.

      "Add this list of twenty data requests to the engagement."

    • atlas_delivery_engagements_info_requests_template_applydelivery:write

      Add every request from an information request template to the engagement, with due dates counted from a start date. Returns the created requests.

      "Apply the financial due diligence request template starting next Monday."

    • atlas_delivery_engagements_info_requests_overdue_listdelivery:read

      List information requests past their due date, optionally only those that block the plan. Use this to answer "what is the client late on".

      "Which blocking information requests are overdue?"

    • atlas_delivery_engagements_info_requests_summary_getdelivery:read

      Read the information request board: counts by status, overdue and blocking counts, and completion. Use this for a quick picture before reading individual requests.

      "How far along is the client with our information requests?"

    • atlas_delivery_engagements_info_requests_getdelivery:read

      Read one information request in full: what was asked, of whom, its status, due date, checklist progress, reminders and history.

      "Show me the status of the payroll data request."

    • atlas_delivery_engagements_info_requests_updatedelivery:write

      Change the fields of one information request, such as its title, due date, priority, owner or blocking flag. Only the fields sent change; send null to clear a field. Returns the updated request.

      "Push the due date on the contracts request to the end of the month."

    • atlas_delivery_engagements_info_requests_deletedelivery:write

      Delete an information request that was added in error and never sent. A request the client has seen should be cancelled instead. Returns the result of the deletion.

      "Delete the duplicate request I just added."

    • atlas_delivery_engagements_info_requests_attachments_listdelivery:read

      List the files attached to an information request, which is what the client has provided against it.

      "What files has the client uploaded for the tax request?"

    • atlas_delivery_engagements_info_requests_attachments_linkdelivery:write

      Link an attachment already uploaded to Atlas to an information request, as evidence provided against it. Returns the link.

      "Attach the file I uploaded to the bank statements request."

    • atlas_delivery_engagements_info_requests_canceldelivery:write

      Withdraw an information request that is no longer needed, with the reason. The client stops being chased for it. Returns the updated request.

      "Cancel the request for the old org chart, we no longer need it."

    • atlas_delivery_engagements_info_requests_escalatedelivery:write

      Escalate an overdue information request to a higher level, with a note. Returns the updated request.

      "Escalate the overdue data extract request to level 2."

    • atlas_delivery_engagements_info_requests_items_listdelivery:read

      List the checklist items within one information request, with whether each has been received and accepted.

      "Which items in the HR request are still missing?"

    • atlas_delivery_engagements_info_requests_items_adddelivery:write

      Add checklist items to an information request, one per label. Returns the added items.

      "Add "leaver list" and "joiner list" to the HR request."

    • atlas_delivery_engagements_info_requests_items_updatedelivery:write

      Update one checklist item within an information request: its label, whether it was received and from whom, whether it was accepted, and a note. Returns the updated item.

      "Mark the leaver list as received from Anna."

    • atlas_delivery_engagements_info_requests_partial_receipt_recorddelivery:write

      Record that part of an information request has been received, with a note saying what is still missing. Returns the updated request.

      "The client sent two of the three years, mark it partially received."

    • atlas_delivery_engagements_info_requests_not_applicable_markdelivery:write

      Mark an information request as not applicable to this client, with the reason. It leaves the outstanding counts. Returns the updated request.

      "The client has no overseas subsidiaries, mark that request not applicable."

    • atlas_delivery_engagements_info_requests_reminddelivery:write

      Send a reminder to the person an information request is asked of. Returns the updated request with its reminder count.

      "Send a reminder for the payroll data request."

    • atlas_delivery_engagements_info_requests_senddelivery:write

      Send an information request to the client, optionally with a due date and a note. It moves from not started to requested and the reminder schedule begins. Returns the updated request.

      "Send the management accounts request to the client, due on the 14th."

    • atlas_delivery_engagements_info_requests_submitdelivery:write

      Record that the client has submitted an information request, in full or in part, with a note and who submitted it. It then waits for review. Returns the updated request.

      "Record that the finance director submitted the ledger extract."

    • atlas_delivery_engagements_raci_cells_listdelivery:read

      Read the responsibility matrix of an engagement: every cell with its subject, party and letter under the decision model in use. Use this to answer "who is accountable for this deliverable".

      "Who is accountable for each deliverable on this engagement?"

    • atlas_delivery_engagements_raci_cells_setdelivery:write

      Set one cell of the responsibility matrix: a party and a letter against a subject. An existing cell for the same subject and party is replaced. Returns the cell.

      "Make Priya accountable for the final report."

    • atlas_delivery_engagements_raci_cells_bulk_setdelivery:write

      Set up to two hundred cells of the responsibility matrix in one request. Either every cell is written or none is. Returns the cells.

      "Fill in the matrix for all six workstreams from this table."

    • atlas_delivery_engagements_raci_cells_updatedelivery:write

      Change the letter or the note on one responsibility matrix cell. Returns the updated cell.

      "Change Sam from consulted to informed on the data model."

    • atlas_delivery_engagements_raci_cells_deletedelivery:write

      Remove one cell from the responsibility matrix. Returns the result of the removal.

      "Remove Tom from the pricing deliverable in the matrix."

    • atlas_delivery_engagements_raci_template_applydelivery:write

      Apply a responsibility matrix template to the engagement, and optionally fill the named seats from the engagement team. Returns the resulting cells.

      "Apply the standard transformation RACI template and fill the seats."

    • atlas_delivery_engagements_raci_exportdelivery:read

      Render the responsibility matrix as a Markdown document, for the internal team or in the client-safe form. Returns the Markdown text, or a job ticket when the matrix is too large to render in one request.

      "Give me the RACI as a document I can paste into the kickoff pack."

    • atlas_delivery_engagements_raci_parties_listdelivery:read

      List every party that appears in the responsibility matrix, with the key used to remove one. Use this to answer "who is on the matrix".

      "Who appears on the RACI for this engagement?"

    • atlas_delivery_engagements_raci_parties_removedelivery:write

      Take one party off the whole responsibility matrix, removing every cell they hold. Use this when a person leaves the engagement. Returns what was removed.

      "Take Tom off the RACI entirely, he has rolled off."

    • atlas_delivery_engagements_raci_subjects_listdelivery:read

      List the engagement records that can appear as rows of the responsibility matrix: phases, workstreams, deliverables and the other subjects, with their ids.

      "What can I assign responsibilities against on this engagement?"

    • atlas_delivery_engagements_raci_decision_model_setdelivery:write

      Switch the decision model the responsibility matrix uses: RACI, RASCI, DACI or RAPID. Returns the matrix under the new model.

      "Switch this engagement to DACI."

    • atlas_delivery_engagements_raci_validatedelivery:read

      Check the responsibility matrix for structural problems, such as a subject with no one accountable or with more than one. Returns the findings; an empty list means none were found.

      "Is anything missing an accountable owner in the RACI?"

    • atlas_delivery_engagements_responsibility_areas_listdelivery:read

      List the responsibility areas of an engagement, in order, with who holds each seat.

      "What responsibility areas are defined on this engagement?"

    • atlas_delivery_engagements_responsibility_areas_createdelivery:write

      Add a responsibility area to an engagement, with a stable key and a label. Returns the created area.

      "Add a responsibility area for data protection."

    • atlas_delivery_engagements_responsibility_areas_updatedelivery:write

      Rename a responsibility area. The key stays the same, so references to it are unaffected. Returns the updated area.

      "Rename the security area to information security."

    • atlas_delivery_engagements_responsibility_areas_deletedelivery:write

      Remove a responsibility area from an engagement. Returns the result of the removal.

      "Remove the procurement responsibility area."

    • atlas_delivery_engagements_responsibility_areas_assigndelivery:write

      Fill the named seats of the responsibility areas from the engagement team, following the playbook. Adding a person to the team never does this on its own, so nobody becomes accountable by accident. Returns the areas with their seats.

      "Fill the responsibility seats from the team we have staffed."

    • atlas_delivery_engagements_responsibility_areas_reorderdelivery:write

      Set the order of the responsibility areas, by sending every area id first to last. Returns the areas in their new order.

      "Put the data protection area first."

    • atlas_delivery_engagements_escalation_paths_listdelivery:read

      List the escalation paths of an engagement: each level with its description, trigger, response time and the firm and client people at that level.

      "Who do we escalate to at level 2 on this engagement?"

    • atlas_delivery_engagements_escalation_paths_createdelivery:write

      Add a level to the escalation path of an engagement, with its trigger, response time, and the firm and client people at that level. Returns the created level.

      "Add a level 3 escalation to the client CFO with a 24 hour response."

    • atlas_delivery_engagements_ways_of_working_getdelivery:read

      Read the ways-of-working charter of an engagement: working hours, meeting cadence, channels, response times, conventions, definitions of done and ready, and whether the client has agreed it.

      "What did we agree with the client about response times?"

    • atlas_delivery_engagements_ways_of_working_setdelivery:write

      Write the ways-of-working charter of an engagement. Only the fields sent change. Returns the charter.

      "Record that core hours are 10 to 4 UK time and responses are due within one working day."

    • atlas_delivery_engagements_operating_model_getdelivery:read

      Describe the operating model of an engagement: the governance forums, the cadence, the roles and the controls that apply to it, as configured.

      "How is this engagement governed?"

    • atlas_delivery_engagements_follow_ups_listdelivery:read

      List the follow-ups on an engagement: reminders and chases with their assignee, subject, due time, recurrence and state.

      "What follow-ups are set on this engagement?"

    • atlas_delivery_engagements_follow_ups_createdelivery:write

      Create a follow-up on an engagement: a reminder or chase for a person, about a record, due at a time, optionally repeating and escalating after missed reminders. Returns the created follow-up.

      "Remind the client sponsor every Monday to confirm the steering attendees."

    • atlas_delivery_engagements_follow_ups_due_listdelivery:read

      List follow-ups falling due within the given number of days. Use this to answer "what needs chasing this week".

      "What follow-ups are due in the next three days?"

    • atlas_delivery_engagements_follow_ups_mine_listdelivery:read

      List the follow-ups on an engagement that are assigned to the person the token belongs to.

      "What follow-ups do I own on this engagement?"

    • atlas_delivery_engagements_follow_ups_getdelivery:read

      Read one follow-up in full, including its schedule, recurrence, channels and escalation settings.

      "Show me the steering attendees reminder."

    • atlas_delivery_engagements_follow_ups_updatedelivery:write

      Change a follow-up: its title, body, subject label, due time, lead time, channels or escalation. Only the fields sent change; send null to clear a field. Returns the updated follow-up.

      "Move the invoice chase to Thursday."

    • atlas_delivery_engagements_follow_ups_canceldelivery:write

      Cancel a follow-up so no further reminders are sent. Returns the cancelled follow-up.

      "Cancel the reminder about the kickoff pack."

    • atlas_delivery_engagements_follow_ups_acknowledgedelivery:write

      Acknowledge a follow-up, recording that the assignee has seen it. Reminders continue until it is completed. Returns the updated follow-up.

      "Acknowledge the reminder about the status report."

    • atlas_delivery_engagements_follow_ups_completedelivery:write

      Complete a follow-up. A repeating follow-up schedules its next occurrence. Returns the updated follow-up.

      "Mark the chase for the signed SOW as done."

    • atlas_delivery_engagements_follow_ups_pausedelivery:write

      Pause a follow-up so no reminders are sent until it is resumed. Returns the updated follow-up.

      "Pause the weekly reminder while the client is on holiday."

    • atlas_delivery_engagements_follow_ups_resumedelivery:write

      Resume a paused follow-up, so its reminders are sent again on schedule. Returns the updated follow-up.

      "Resume the weekly reminder."

    • atlas_delivery_engagements_follow_ups_runs_listdelivery:read

      Read the delivery history of one follow-up: every reminder sent, on which channel, and whether it was acknowledged or escalated.

      "Has the client sponsor been reminded about this yet?"

    • atlas_delivery_engagements_follow_ups_snoozedelivery:write

      Snooze a follow-up until a given time. Returns the updated follow-up.

      "Snooze the invoice chase until next Tuesday."

    • atlas_delivery_engagements_analyses_listdelivery:read

      List the analyses on an engagement, optionally by status, hypothesis or workstream. Each carries its type, owner, dates and findings.

      "Which analyses are blocked?"

    • atlas_delivery_engagements_analyses_createdelivery:write

      Plan an analysis on an engagement: what it is, which hypothesis it tests, its type, owner, reviewer and dates. Returns the created analysis.

      "Plan a customer churn analysis to test hypothesis 2, owned by Sam, due in two weeks."

    • atlas_delivery_engagements_analyses_updatedelivery:write

      Change the fields of an analysis, including its findings, so-what and confidence while it is in progress. Only the fields sent change. Returns the updated analysis.

      "Update the churn analysis finding to say retention is worst in the SME segment."

    • atlas_delivery_engagements_analyses_startdelivery:write

      Start a planned analysis, moving it from planned to in progress. Returns the updated analysis.

      "Start the pricing benchmark analysis."

    • atlas_delivery_engagements_analyses_blocked_markdelivery:write

      Mark an analysis as blocked, with the reason. Returns the updated analysis.

      "The churn analysis is blocked waiting on the CRM extract."

    • atlas_delivery_engagements_analyses_completedelivery:write

      Complete an analysis with its finding summary and so-what, and optionally its confidence and actual effort. Only a completed analysis counts as evidence behind a storyline page. Returns the updated analysis.

      "Complete the churn analysis: churn is concentrated in year-one SME customers."

    • atlas_delivery_engagements_hypotheses_listdelivery:read

      List the questions and hypotheses on an engagement, optionally by kind, status or tree path, flat rather than as a tree.

      "Which hypotheses are still being tested?"

    • atlas_delivery_engagements_hypotheses_createdelivery:write

      Add a question or hypothesis to the engagement tree, optionally under a parent. Returns the created hypothesis.

      "Add a hypothesis under the growth question: pricing is below the market."

    • atlas_delivery_engagements_hypotheses_updatedelivery:write

      Change the statement, rationale, owner, confidence or other fields of a hypothesis. Only the fields sent change; send null to clear a field that allows it. Returns the updated hypothesis.

      "Raise the confidence on the pricing hypothesis to 70."

    • atlas_delivery_engagements_hypotheses_resolvedelivery:write

      Conclude a hypothesis as supported, partially supported, refuted or inconclusive, with the conclusion. A concluded leaf with no completed analysis behind it is flagged in the tree. Returns the updated hypothesis.

      "Conclude the pricing hypothesis as supported."

    • atlas_delivery_engagements_hypotheses_movedelivery:write

      Move a hypothesis to a different parent, or to the top level with a null parent, at a position among its siblings. Returns the updated hypothesis.

      "Move the channel hypothesis under the growth question."

    • atlas_delivery_engagements_interviews_listdelivery:read

      List the interviews on an engagement, optionally by status or interviewee category.

      "Which customer interviews are still to be scheduled?"

    • atlas_delivery_engagements_interviews_createdelivery:write

      Add an interview to the engagement programme: who, their organisation and category, when, and who interviews them. Returns the created interview.

      "Add an interview with the head of procurement at the client, next Tuesday."

    • atlas_delivery_engagements_interviews_completedelivery:write

      Record that an interview took place, with key quotes, a finding summary, the interviewee sentiment and whether a follow-up is needed. Returns the updated interview.

      "Record the procurement interview: supportive, main concern is supplier lock-in."

    • atlas_delivery_engagements_interviews_consent_recorddelivery:write

      Record the consent the interviewee gave: whether consent was obtained, on what basis, and whether recording and attribution are permitted. Returns the updated interview.

      "Record that the interviewee consented but not to be quoted by name."

    • atlas_delivery_engagements_sources_listdelivery:read

      List the sources on an engagement with their kind, reliability grade and credibility.

      "Which sources have we used and how reliable are they?"

    • atlas_delivery_engagements_sources_createdelivery:write

      Add a source to an engagement: its title, kind, citation, dates, reliability, licence terms and whether it may be redistributed. Returns the created source.

      "Add the 2025 industry market report as a source."

    • atlas_delivery_engagements_sources_ratedelivery:write

      Grade a source on the reliability scale from A to F and the credibility scale from 1 to 6, and record a retraction when there is one. Returns the updated source.

      "Grade the market report B2."

    • atlas_delivery_engagements_sources_triangulation_recorddelivery:write

      Record which other sources corroborate or contradict a source. Returns the updated source.

      "The survey corroborates the market report and the vendor figures contradict it."

    • atlas_delivery_engagements_storyline_sections_createdelivery:write

      Add a section to the storyline: a page title, governing thought, section type, and the analysis and deliverable behind it. Returns the created section.

      "Add a storyline page: SME churn is the biggest value lever."

    • atlas_delivery_engagements_storyline_sections_updatedelivery:write

      Change a storyline section, including its status between ghost, draft and final. Only the fields sent change. Returns the updated section.

      "Move the churn page to draft."

    • atlas_delivery_engagements_storyline_sections_reorderdelivery:write

      Set the order of storyline sections under one parent, by sending every section id first to last. Returns the sections in their new order.

      "Put the recommendation page before the next steps page."

    • atlas_delivery_engagements_storyline_sections_verifydelivery:write

      Mark a storyline section as evidenced. The API refuses unless the page has a governing thought and a completed analysis behind it, and names what is missing. Returns the updated section.

      "Mark the churn page as evidenced."

    • atlas_delivery_engagements_working_paper_file_getdelivery:read

      Read the working paper file of an engagement: its assembly standard, report date, retention, lock state, legal holds and completeness.

      "Is the working paper file for this engagement locked?"

    • atlas_delivery_engagements_working_paper_file_createdelivery:write

      Open the working paper file of an engagement under an assembly standard, with its structure template, report date, retention period and storage jurisdiction. Returns the file.

      "Open the working paper file under ISA 230 with seven years retention."

    • atlas_delivery_engagements_working_paper_file_updatedelivery:write

      Change the settings of an unlocked working paper file, such as its report date, retention or completeness checklist status. Returns the updated file.

      "Set the report date on the working paper file to 31 March."

    • atlas_delivery_engagements_working_paper_file_readiness_getdelivery:read

      Check whether the working paper file is ready to lock: papers not prepared or reviewed, review notes still open, and the assembly deadline. Returns the blockers; an empty list means it is ready.

      "What is stopping us from locking the working papers?"

    • atlas_delivery_engagements_working_paper_file_verifydelivery:read

      Verify the integrity of a locked working paper file against its recorded fingerprints. Returns whether every paper is unchanged since the lock, and which are not.

      "Has anything in the locked file changed since it was locked?"

    • atlas_delivery_engagements_working_paper_holds_createdelivery:write

      Place a legal hold on the working paper file, with the reason. A held file cannot be destroyed. Returns the updated file.

      "Put a legal hold on the working papers because of the regulator enquiry."

    • atlas_delivery_engagements_working_papers_listdelivery:read

      List the papers in the working paper file with their section, index reference, preparer and reviewer sign-offs.

      "Which working papers are still unreviewed?"

    • atlas_delivery_engagements_working_papers_createdelivery:write

      Add a paper to the working paper file under a section, with its purpose, index reference, attachment and cross-references. Adding to a locked file needs a reason and an approver. Returns the created paper.

      "Add the revenue testing paper under section C."

    • atlas_delivery_engagements_models_listdelivery:read

      List the models registered on an engagement with their type, tier, validation status and owner.

      "Which models are on this engagement?"

    • atlas_delivery_engagements_models_createdelivery:write

      Register a model built on the engagement: its type, purpose, tier, owners, inputs, assumptions, controls, limitations and, for a model based on a large language model, its provider and controls. Returns the registered model.

      "Register the pricing elasticity model as a tier 2 statistical model owned by Sam."

    • atlas_delivery_engagements_models_unvalidated_listdelivery:read

      List the models that have not passed independent validation. Use this to answer "which models are in use without validation".

      "Which models have not been validated yet?"

    • atlas_delivery_engagements_models_getdelivery:read

      Read one registered model in full, including its validation findings and approved purposes.

      "Show me the pricing model entry."

    • atlas_delivery_engagements_models_updatedelivery:write

      Change the fields of a registered model. Only the fields sent change; send null to clear a person field. Returns the updated model.

      "Record that sensitivity testing has been done on the pricing model."

    • atlas_delivery_engagements_models_archivedelivery:write

      Retire a model so it is no longer used, with the reason and the model that replaces it. Returns the updated model.

      "Retire version 1 of the pricing model, replaced by version 2."

    • atlas_delivery_engagements_analytical_assumptions_listdelivery:read

      List the analytical assumptions of an engagement: each variable with its base, low and high case, basis, validation status and which models use it.

      "What assumptions drive the business case?"

    • atlas_delivery_engagements_analytical_assumptions_createdelivery:write

      Record an analytical assumption: the statement, the variable, its basis, its base, low and high values, and where it is used. Returns the recorded assumption.

      "Record the assumption that churn stays at 8 percent, based on client data."

    • atlas_delivery_engagements_analytical_assumptions_updatedelivery:write

      Revise the basis, disclosure deliverable or expiry of an analytical assumption. Returns the updated assumption.

      "Change the basis of the churn assumption to benchmark."

    • atlas_delivery_engagements_analytical_assumptions_voiddelivery:write

      Record that an analytical assumption no longer holds, with a note, and whether figures that rely on it must be restated. Returns the updated assumption.

      "The churn assumption is wrong, invalidate it and flag a restatement."

    • atlas_delivery_engagements_benchmarks_listdelivery:read

      List the benchmarks on an engagement, optionally only those without a competition law review or those attached to one deliverable.

      "Which benchmarks still need a competition law review?"

    • atlas_delivery_engagements_benchmarks_createdelivery:write

      Record a benchmark: the metric and its definition, the peer set, the data source and vintage, the quartiles, the client value and gaps, and licence and anonymisation constraints. Returns the recorded benchmark.

      "Record the cost-to-serve benchmark against the 12 peer banks."

    • atlas_delivery_engagements_benchmarks_updatedelivery:write

      Change the fields of a benchmark. Only the fields sent change; send null to clear a field that allows it. Returns the updated benchmark.

      "Update the client value on the cost-to-serve benchmark to 41.2."

    • atlas_delivery_engagements_benchmarks_deletedelivery:write

      Delete a benchmark recorded in error. Returns the result of the deletion.

      "Delete the duplicate benchmark."

    • atlas_delivery_engagements_benchmarks_linkdelivery:write

      Attach a benchmark to a deliverable, so it is shown and checked as part of it. Returns the updated benchmark.

      "Attach the cost-to-serve benchmark to the final report."

    • atlas_delivery_engagements_benchmarks_unlinkdelivery:write

      Detach a benchmark from the deliverable it is attached to. Returns the updated benchmark.

      "Take the cost-to-serve benchmark off the final report."

    • atlas_delivery_engagements_expert_calls_listdelivery:read

      List the expert network calls on an engagement with their network, topic, compliance state and status.

      "Which expert calls are waiting for compliance approval?"

    • atlas_delivery_engagements_expert_calls_requestdelivery:write

      Request an expert network call: the network, expert reference, topic, hypotheses it tests, justification, the expert role history and compliance flags. It then waits for compliance review and approval. Returns the call.

      "Request a call with a former procurement director from the GLG network about supplier pricing."

    • atlas_delivery_engagements_expert_calls_incidents_listdelivery:read

      List expert calls with an open compliance incident, such as a call terminated early for sharing restricted information.

      "Are there any open expert call incidents?"

    • atlas_delivery_engagements_expert_calls_getdelivery:read

      Read one expert call in full, including its compliance review, attestation and incident record.

      "Show me the supplier pricing expert call."

    • atlas_delivery_engagements_expert_calls_updatedelivery:write

      Change an expert call: its topic, justification, employer history, compliance flags, chaperone, attestation and script status, or notes attachment. Only the fields sent change. Returns the updated call.

      "Record that the expert attestation has been obtained."

    • atlas_delivery_engagements_expert_calls_completedelivery:write

      Record that an approved expert call took place, with its duration, attendees and notes attachment. Returns the updated call.

      "Record that the expert call happened, 45 minutes, with Sam and Priya."

    • atlas_delivery_engagements_expert_calls_incidents_createdelivery:write

      Flag a compliance incident on an expert call, such as a call ended early because the expert began to share restricted information. Returns the updated call.

      "Flag an incident on the call: it was stopped when the expert mentioned current pricing."

    • atlas_delivery_engagements_document_handling_listdelivery:read

      List the handling rules recorded for the documents of an engagement: classification, markings, distribution limits, redaction and metadata state, inside information flags, and release state.

      "What are the handling rules on the board paper?"

    • atlas_delivery_engagements_document_handling_setdelivery:write

      Record or update the handling rules for a document: classification, markings, who may receive it, print and download permissions, DRM, redaction and metadata scrubbing, and inside information flags. Returns the record.

      "Mark the board paper as named recipients only with no download."

    • atlas_delivery_engagements_document_handling_withdrawdelivery:write

      Recall a released document, with the reason, so recipients are told to stop using it. Returns the updated record.

      "Recall the board paper, it went out with the wrong figures."

    • atlas_delivery_engagements_data_dispositions_listdelivery:read

      List the data dispositions of an engagement: each client dataset, where it is held, what the client instructed, and whether it has been returned or destroyed.

      "What client data do we still hold for this engagement?"

    • atlas_delivery_engagements_data_dispositions_createdelivery:write

      Record a client dataset and the client instruction for it at the end of the engagement: its classification, locations, return format and any conflict with the firm retention obligations. Returns the record.

      "Record that the client wants the customer extract destroyed, held in two locations."

    • atlas_delivery_engagements_data_dispositions_outstanding_listdelivery:read

      List the data dispositions not yet carried out. Use this to answer "what client data must still be returned or destroyed".

      "Which client datasets have not been returned or destroyed yet?"

    • atlas_delivery_engagements_threads_createdelivery:write

      Open a thread on an engagement: a question, clarification, request, blocker, review comment, escalation or complaint, asked of a named party, with a response time. Returns the created thread.

      "Ask the client IT lead whether we can get VPN access by Friday."

    • atlas_delivery_engagements_threads_first_response_recorddelivery:write

      Record the first response on a thread, which stops its first-response clock. Returns the updated thread.

      "Record that we have responded to the client question on scope."

    • atlas_delivery_engagements_threads_answerdelivery:write

      Record the answer to a thread, with a summary. Returns the updated thread.

      "Record the answer: VPN access will be ready on Thursday."

    • atlas_delivery_engagements_threads_closedelivery:write

      Close a thread, with the reason, whether or not it was answered. Returns the updated thread.

      "Close the VPN thread, access is working."

    • atlas_delivery_engagements_threads_reopendelivery:write

      Reopen a closed or answered thread, so its response clock runs again. Returns the updated thread.

      "Reopen the VPN thread, access has broken again."

    • atlas_delivery_engagements_feedback_listdelivery:read

      List the feedback requests and responses on an engagement: kind, respondent, score, verbatim and themes.

      "What feedback has the client given on this engagement?"

    • atlas_delivery_engagements_feedback_requestdelivery:write

      Ask for feedback on an engagement: the kind of survey, the respondent, the themes and the milestone that triggered it. Returns the feedback request.

      "Ask the client sponsor for an NPS score after phase one."

    • atlas_delivery_engagements_feedback_follow_ups_listdelivery:read

      List feedback responses that need a follow-up conversation, such as a low score, and who owns each.

      "Which feedback responses still need a follow-up call?"

    • atlas_delivery_engagements_feedback_quotable_listdelivery:read

      List feedback the respondent has consented to have quoted, for use in proposals and case studies.

      "Which client quotes are we allowed to use?"

    • atlas_delivery_engagements_feedback_responses_recorddelivery:write

      Record the response to a feedback request: the score, verbatim comment, themes and the owner of any follow-up. Returns the updated feedback.

      "Record a score of 9 from the sponsor with the comment "excellent team"."

    • atlas_delivery_engagements_feedback_consent_recorddelivery:write

      Record whether the respondent consents to be quoted and to act as a reference. Returns the updated feedback.

      "The sponsor is happy to be quoted but not to be a reference."

    • atlas_delivery_engagements_feedback_follow_ups_completedelivery:write

      Record that the follow-up conversation on a feedback response has happened. Returns the updated feedback.

      "Mark the follow-up call on the low score as done."

    • atlas_delivery_engagements_complaints_listdelivery:read

      List the client complaints on an engagement with their severity, category, owner, status and target resolution date.

      "Are there any open complaints on this engagement?"

    • atlas_delivery_engagements_complaints_createdelivery:write

      Record a client complaint: its severity, category, description, who raised it, the owner and the target resolution date. Returns the complaint.

      "Log a high severity complaint from the CFO about late reports."

    • atlas_delivery_engagements_complaints_acknowledgedelivery:write

      Acknowledge a complaint, which stops its acknowledgement clock. Returns the updated complaint.

      "Acknowledge the CFO complaint."

    • atlas_delivery_engagements_complaints_escalatedelivery:write

      Escalate a complaint to the next level of the firm. Returns the updated complaint.

      "Escalate the CFO complaint."

    • atlas_delivery_engagements_complaints_closedelivery:write

      Close a resolved complaint with its root cause and preventive action, optionally linked to the RAID item that records the underlying problem. Returns the updated complaint.

      "Close the late reports complaint: root cause was a missing review step."

    • atlas_delivery_engagements_lessons_listdelivery:read

      List the lessons learned on an engagement with their category, severity, applicability and whether they have been published.

      "What lessons have we captured on this engagement?"

    • atlas_delivery_engagements_lessons_createdelivery:write

      Capture a lesson learned: its category, the situation, what happened, the impact, the recommendation, and how widely it applies. Returns the lesson.

      "Capture a lesson: we under-estimated data cleansing effort by half."

    • atlas_delivery_engagements_benefits_listdelivery:read

      List the benefits register of an engagement: each benefit with its type, measure, baseline, target, attribution and readings.

      "What benefits are we tracking for the client?"

    • atlas_delivery_engagements_benefits_createdelivery:write

      Add a benefit to the register: its type, name, measure, unit, baseline and target values, attribution percentage and currency. Returns the benefit.

      "Track a cost reduction benefit of 2 million pounds against a 10 million baseline."

    • atlas_delivery_engagements_benefits_readings_recorddelivery:write

      Record a reading of a benefit at a date, with its source. Returns the reading.

      "Record a reading of 1.2 million on the cost benefit as of 30 June."

    • atlas_delivery_engagements_assets_listdelivery:read

      List the reusable assets harvested from an engagement with their kind, classification, consent and sanitisation state.

      "What reusable assets came out of this engagement?"

    • atlas_delivery_engagements_assets_createdelivery:write

      Harvest a reusable asset from an engagement: its kind, title, classification, tags, attachment, and whether client consent is required before reuse. Returns the asset.

      "Harvest the pricing model template as a reusable asset."

    • atlas_delivery_engagements_assets_consent_recorddelivery:write

      Record that the client has consented to reuse of an asset, with the evidence. Returns the updated asset.

      "Record the client consent to reuse the case study, evidence attached."

    • atlas_delivery_engagements_assets_reuse_recorddelivery:write

      Record that an asset has been reused. The API refuses reuse of an asset that still needs consent or sanitisation. Returns the updated asset.

      "Record that we reused the pricing template."

    • atlas_delivery_engagements_handoff_information_requestdelivery:write

      Ask sales for more information on a submitted handoff, with the detail needed. Returns the updated handoff.

      "Ask sales to clarify what "light touch PMO" means in the handoff."

    • atlas_delivery_engagements_transition_plan_getdelivery:read

      Read the transition plan of an engagement: who takes over, hypercare and warranty windows, support model, knowledge transfer, residual risks, staff transfer obligations and the rollback plan.

      "When does hypercare end on this engagement?"

    • atlas_delivery_engagements_transition_plan_setdelivery:write

      Write the transition plan of an engagement. The receiving team is required; every other field is optional. Returns the plan.

      "Set the transition plan: the client service desk takes over with four weeks of hypercare."

    • atlas_delivery_engagements_staffing_requests_listdelivery:read

      List the staffing requests of an engagement: role, grade, skills, allocation, needed-by date, proposed person and status.

      "Which roles on this engagement are still unfilled?"

    • atlas_delivery_engagements_staffing_requests_createdelivery:write

      Raise a staffing request for a role on an engagement: its grade, skills, allocation, needed-by date, whether it is key personnel, and whether the client must approve the person. Returns the request.

      "Raise a request for a senior data engineer at 100 percent from 1 March."

    • atlas_delivery_engagements_staffing_requests_recommenddelivery:write

      Propose a person for a staffing request. Returns the updated request.

      "Propose Priya for the data engineer role."

    • atlas_delivery_engagements_staffing_requests_allocations_listdelivery:read

      Read the effective allocation of every staffed person in a given engagement week, after their ramp profile is applied.

      "What is the effective team capacity in week 3?"

    • atlas_delivery_engagements_ramp_profiles_listdelivery:read

      List the ramp profiles of an engagement: how productivity builds week by week for a joiner, and the handover and ramp-down terms.

      "What ramp profiles are defined on this engagement?"

    • atlas_delivery_engagements_ramp_profiles_createdelivery:write

      Define a ramp profile: the curve type, productivity by week, onboarding days that are not billable, learning and shadowing hours, handover overlap and ramp-down taper. Returns the profile.

      "Define a ramp where a joiner is at 50, 75 then 100 percent over three weeks."

    • atlas_delivery_engagements_ramp_profiles_allocation_calculatedelivery:read

      Calculate the effective allocation of a planned allocation in a given week under a ramp profile.

      "What is a full-time joiner actually worth in week 2 under this ramp?"

    • atlas_delivery_engagements_procurement_getdelivery:read

      Read the client procurement record of an engagement: supplier registration, purchase order number, value, consumption and expiry, invoicing portal and instructions, payment terms and billing contact.

      "How much of the purchase order is left?"

    • atlas_delivery_engagements_procurement_setdelivery:write

      Write the client procurement record of an engagement. Only the fields sent change. Returns the record.

      "Record PO 4500123 for 250,000 GBP expiring at the end of the year."

    • atlas_delivery_engagements_procurement_consumption_recorddelivery:write

      Record consumption against the purchase order, such as an invoice raised against it. Returns the record with the remaining value.

      "Record 40,000 consumed against the PO for the March invoice."

    • atlas_delivery_engagements_pricing_assumptions_listdelivery:read

      List the pricing assumptions the engagement was sold on, with what the sale assumed, what is true now, and any breach recorded.

      "Which pricing assumptions have been breached?"

    • atlas_delivery_engagements_pricing_assumptions_createdelivery:write

      Record a pricing assumption the engagement was sold on, with the sold basis and the current reality. Returns the assumption.

      "Record that pricing assumed the client provides test environments."

    • atlas_delivery_engagements_pricing_assumptions_breach_recorddelivery:write

      Record that a pricing assumption has been breached: what is true now, the commercial impact, and the change request that covers it. Returns the updated assumption.

      "The client cannot provide test environments, record the breach against CR-12."

    • atlas_delivery_engagements_commercial_terms_listdelivery:read

      List the commercial terms recorded from the contract of an engagement: liability caps, intellectual property, warranties, service levels, key personnel, acceptance, audit, benchmarking and others.

      "What is the liability cap on this engagement?"

    • atlas_delivery_engagements_commercial_terms_createdelivery:write

      Record a commercial term from the contract. Send the term type and the fields that apply to it: caps, intellectual property, warranty, service levels, key personnel, acceptance, audit, step-in, benchmarking and so on. Returns the term.

      "Record a liability cap of 1 times fees paid over a rolling 12 months."

    • atlas_delivery_engagements_commercial_terms_deletedelivery:write

      Delete a commercial term recorded in error. Returns the result of the deletion.

      "Delete the duplicate warranty term."

    • atlas_delivery_engagements_fee_caps_calculatedelivery:read

      Calculate the audit independence fee caps for an engagement: the non-audit fee cap against the three-year audit fee average and the fee dependency threshold. Use this before taking on non-audit work for an audit client.

      "How much non-audit work can we take on for this audit client?"

    • atlas_delivery_engagements_ai_drafts_listdelivery:read

      List the AI drafts proposed on an engagement, optionally by kind and status. A draft changes nothing until a person accepts it.

      "Which AI drafts are waiting for review?"

    • atlas_delivery_engagements_ai_drafts_getdelivery:read

      Read one AI draft with its proposed content and the keys that can be accepted.

      "Show me the draft minutes for the steering meeting."

    • atlas_delivery_engagements_ai_drafts_rejectdelivery:write

      Reject an AI draft, with an optional note, so it is not applied. Returns the updated draft.

      "Reject the draft status report, the tone is wrong."

    • atlas_delivery_engagements_ai_drafts_deliverable_qa_generatedelivery:write

      Ask for an AI quality pre-check of a deliverable. The result is a proposed draft of findings that a person reviews. Returns the draft.

      "Pre-check the final report before it goes to review."

    • atlas_delivery_engagements_ai_drafts_health_explanation_generatedelivery:write

      Ask for an AI explanation of the current health ratings of an engagement, drawn from its registers. Returns the proposed draft.

      "Explain why this engagement is amber."

    • atlas_delivery_engagements_ai_drafts_minutes_generatedelivery:write

      Ask for AI draft minutes of a governed meeting. The minutes are a proposal until a person accepts them. Returns the draft.

      "Draft the minutes for yesterday's steering meeting."

    • atlas_delivery_engagements_ai_drafts_raid_items_generatedelivery:write

      Ask the AI to propose RAID items from up to ten meetings and twenty threads. The proposals are a draft until a person accepts them. Returns the draft.

      "Suggest risks from the last three steering meetings."

    • atlas_delivery_engagements_ai_drafts_status_report_generatedelivery:write

      Ask for an AI draft of the narrative of a status report. The narrative is a proposal until a person accepts it. Returns the draft.

      "Draft the narrative for this week's status report."

    • atlas_delivery_engagements_exports_listdelivery:read

      List the export artifacts an engagement offers, with the formats each supports and whether it has a client variant. The economics summary is left out, because it carries cost rates and margin.

      "What documents can I export from this engagement?"

    • atlas_delivery_engagements_exports_renderdelivery:read

      Render one export artifact as a Markdown document, for the internal team or in the client-safe form. Returns the Markdown text, or a job ticket when the artifact is too large to render in one request. For PDF, Word or Excel, use atlas_delivery_engagements_export_jobs_create. The economics summary is refused, because it carries cost rates and margin.

      "Give me the stakeholder map as Markdown."

    • atlas_delivery_engagements_export_jobs_listdelivery:read

      List the export jobs the person has requested on this engagement, newest first, with their status.

      "Are my exports ready?"

    • atlas_delivery_engagements_export_jobs_createdelivery:write

      Request an export artifact as a background job in PDF, Word, Excel or Markdown. Returns the job; poll it with atlas_delivery_engagements_export_jobs_get and fetch it with atlas_delivery_engagements_export_jobs_download. The economics summary is refused, because it carries cost rates and margin.

      "Export the RAID log as a PDF."

    • atlas_delivery_engagements_export_jobs_getdelivery:read

      Read the status of one export job: queued, running, ready or failed, and when a ready artifact expires. Use this to poll a job before downloading it.

      "Is the RAID PDF ready?"

    • atlas_delivery_engagements_export_jobs_downloaddelivery:read

      Issue the one-time download link for a ready export job. The link is spent when issued, so a second call returns 410 and a fresh export is needed. The job is read first, and a job for the economics summary is refused, because it carries cost rates and margin.

      "Give me the link to download the RAID PDF."

    • atlas_delivery_engagements_barriers_listdelivery:read

      List the information barriers on an engagement: each barrier type, reason, the parties on each side, the excluded parties, the enforcement points, any code name and the dates it is in force. Use this to answer "is there an ethical wall on this engagement".

      "Is there an ethical wall on the Northwind engagement?"

    • atlas_delivery_engagements_barriers_crossings_listdelivery:read

      List the open wall crossings on an engagement: who was brought over an information barrier, for what purpose, on whose approval and until when. Closed crossings are not included, because they need no action.

      "Who has been brought over the wall on Project Falcon, and until when?"

    • atlas_delivery_engagements_barriers_crossings_closedelivery:write

      Close an open wall crossing, so the person it admitted is outside the information barrier again. This restores the wall; it cannot admit anybody. Returns the crossing id and whether it was closed.

      "Close the wall crossing for Sam on Project Falcon now that the call is over."

    • atlas_delivery_engagements_barriers_name_getdelivery:read

      Ask which name the caller may use for an engagement under a code-named barrier. Pass the real name; the answer is the real name for a person inside the wall and the code name for anybody else, with a flag saying which it is. Use this before writing about a deal that may be code named.

      "Which name should I use for the Acme deal in my notes?"

    • atlas_delivery_engagements_barriers_insiders_listdelivery:read

      List the insider list of an engagement: every person recorded as holding inside information, why, when they obtained it, and whether they have been notified of their obligations and acknowledged them.

      "Show me the insider list for Project Falcon."

    • atlas_delivery_engagements_barriers_insiders_adddelivery:write

      Add a person to the insider list of an engagement, recording that they hold inside information, their function and the reason, and when they obtained it. This records a fact; it does not admit the person through any barrier. Returns the entry.

      "Add Priya to the Project Falcon insider list; she joined the valuation call today."

    • atlas_delivery_engagements_barriers_insiders_obligation_recorddelivery:write

      Record one step of an insider list obligation: that the person was notified of their duties, or that they acknowledged them. The two are separate steps under market abuse rules, so each is recorded on its own, and only once it has happened. Returns the entry.

      "Record that Priya was notified of her insider obligations."

    • atlas_delivery_engagements_access_grants_listdelivery:read

      List the access grants on an engagement: who holds access, the reason recorded, when it expires and which workstreams it is limited to. Requires administrator access to the engagement.

      "Who has access to the Northwind engagement?"

    • atlas_delivery_engagements_access_grants_revokedelivery:write

      Revoke one access grant on an engagement, so its holder no longer reaches the engagement through it. This removes access and cannot grant any. Returns the revoked grant.

      "Revoke the access grant for the contractor who left the Northwind engagement."

    • atlas_delivery_engagements_access_provisions_listdelivery:read

      List the client system access the team holds on an engagement: each system, access level, team member, when it was granted, whether credentials were issued, when it expires, and whether it has been revoked.

      "Which client systems does the team still have access to on Northwind?"

    • atlas_delivery_engagements_access_provisions_recorddelivery:write

      Record that the client gave a team member access to one of its systems, with the access level, dates and whether credentials were issued. This is the register of access already given; it does not grant access to anything. Returns the record.

      "Log that the client gave Omar read-only access to their SAP system today."

    • atlas_delivery_engagements_access_provisions_revokedelivery:write

      Record that access to a client system has been removed, with the evidence of its removal, such as the client ticket reference. Returns the updated record.

      "Record that Omar's SAP access was removed under ticket INC-4412."

    • atlas_delivery_engagements_portal_invites_listdelivery:read

      List the client portal invitations on an engagement: the CRM contact, access role, who invited them, when it was issued and expires, and whether it is outstanding, accepted, expired or revoked. The invitation tokens are never returned.

      "Which client contacts have outstanding portal invitations on Northwind?"

    • atlas_delivery_engagements_authority_listdelivery:read

      List the delegation of authority that applies to an engagement: for each kind of decision and value band, the approver role or person and whether a second approver is required.

      "What is the delegation of authority on this engagement?"

    • atlas_delivery_engagements_authority_chain_getdelivery:read

      Work out who must approve a decision of a given kind and value on an engagement, from its delegation of authority, including the bands that apply to the whole workspace. Returns the approval chain and how many bands were considered. Use this to answer "who has to sign off a write-off of this size".

      "Who has to approve a 12,500 write-off on Northwind?"

    • atlas_delivery_engagements_regulatory_obligations_listdelivery:read

      List the regulatory obligations on an engagement, such as breach notification and reporting deadlines, with their regime, trigger, due date, owner and status. Filter by status or regime.

      "List the open GDPR obligations on this engagement."

    • atlas_delivery_engagements_regulatory_obligations_due_listdelivery:read

      List the regulatory obligations on an engagement that fall due within a window, together with every one already overdue. Use this to answer "what regulatory deadlines are coming up".

      "Which regulatory deadlines fall due in the next 48 hours?"

    • atlas_delivery_engagements_regulatory_obligations_generatedelivery:write

      Create the standard obligations of one or more regulatory regimes on an engagement from the catalogue, with their deadlines counted from when the triggering event happened. Returns the obligations created.

      "Add the DORA incident reporting obligations, triggered at 09:00 today."

    • atlas_delivery_engagements_regulatory_obligations_createdelivery:write

      Record one regulatory obligation on an engagement that the catalogue does not cover, with its regime, text, trigger, deadline, evidence requirement and owner. Returns the obligation.

      "Record a 72-hour notification obligation to the regulator, starting from the incident this morning."

    • atlas_delivery_engagements_regulatory_obligations_transitiondelivery:write

      Move a regulatory obligation between open and in progress, optionally attaching evidence. Whether an obligation was met, breached or waived is a claim about a legal obligation, so a person records it in Atlas. Returns the obligation.

      "Mark the breach notification obligation as in progress; the draft filing is attached."

    • atlas_delivery_engagements_obligations_listdelivery:read

      List the contract obligation register of an engagement: each clause the team must perform, its type, due pattern, next due date, owner, evidence requirement, breach consequence and status.

      "Show the contract obligations register for Northwind."

    • atlas_delivery_engagements_obligations_due_listdelivery:read

      List the contract obligations of an engagement that fall due within a number of days, together with every one already overdue. Use this to answer "which contract commitments are due".

      "Which contract obligations fall due this month?"

    • atlas_delivery_engagements_obligations_createdelivery:write

      Record one obligation from a contract on an engagement: the instrument and clause, what it requires, its type and due pattern, the owner role, the next due date, the evidence it needs and the consequence of breach. Returns the obligation.

      "Record clause 7.2 of the MSA: a monthly SLA report is due on the fifth."

    • atlas_delivery_engagements_legal_instruments_listdelivery:read

      List the legal instruments of an engagement, such as the engagement letter, statement of work, NDA and data processing agreement, with whether each is required, its status, counterparty, signatories, dates and renewal terms.

      "Is the engagement letter executed on Northwind?"

    • atlas_delivery_engagements_screening_runs_listdelivery:read

      List the sanctions and politically exposed person screening runs on an engagement: the subject, the lists screened, the provider, each hit with its disposition and second approval, and whether monitoring is ongoing.

      "Show the sanctions screening runs for this client."

    • atlas_delivery_engagements_screening_runs_recorddelivery:write

      Record the result of a screening run against an acceptance check: the subject, the lists screened, the provider and list version, and every hit it returned. Hits are recorded undisposed; deciding whether a hit is a true match stays with a person. Returns the run.

      "Record the sanctions screen we just ran on Acme Holdings against OFAC and the UN list, with its two hits."

    • atlas_delivery_engagements_beneficial_owners_listdelivery:read

      List the beneficial owners recorded on an engagement: each owner, how they control the client, their ownership share, their screening status and whether a register discrepancy was reported.

      "Who are the beneficial owners recorded for Acme Holdings?"

    • atlas_delivery_engagements_beneficial_owners_adddelivery:write

      Add a beneficial owner of the client to an acceptance check, with how they control it, their ownership share, their details, their screening results and the registry the information came from. Returns the owner.

      "Add Jane Doe as a 30 percent shareholder beneficial owner on the acceptance check."

    • atlas_delivery_engagements_beneficial_owners_discrepancy_recorddelivery:write

      Record that a discrepancy between the beneficial owner the firm found and the public register has been reported to the register. This records the report; it does not send one. Returns the owner.

      "Record that we reported the Companies House discrepancy for Jane Doe."

    • atlas_delivery_engagements_beneficial_owners_assessment_getdelivery:read

      Apply the fifty percent rule to the beneficial owners of an acceptance check: add up the holdings of every owner recorded as blocked and say whether the client is itself blocked. The result is also written onto the latest screening run.

      "Is Acme Holdings blocked under the fifty percent rule?"

    • atlas_delivery_engagements_reliance_letters_listdelivery:read

      List the reliance letters on an engagement: each relying party, the scope of reliance, the caps, indemnities, dates and whether it has been issued.

      "Which reliance letters exist on the Northwind due diligence?"

    • atlas_delivery_engagements_reliance_letters_exposure_getdelivery:read

      Read the liability exposure of an engagement across every reliance letter it has issued: the caps to each party, the aggregate cap, and any letter with no aggregate cap, which is unlimited.

      "What is our total liability exposure from reliance letters on Northwind?"

    • atlas_delivery_engagements_reliance_letters_draftdelivery:write

      Record a draft reliance letter on an engagement: the relying party, its relationship, the scope and deliverables relied on, the liability caps, fee, indemnities and dates. The letter is recorded unissued; issuing it stays with a person. Money is a decimal string. Returns the letter.

      "Draft a reliance letter for the lender with a 1,000,000 cap."

    • atlas_delivery_engagements_third_party_reliance_listdelivery:read

      List the work of other specialists the engagement relies on: each specialist, the assessments of their competence and objectivity, the firm evaluation of their work, and whether reliance on it has been accepted.

      "Which specialists are we relying on for the Northwind audit?"

    • atlas_delivery_engagements_third_party_reliance_recorddelivery:write

      Record work by another specialist that the engagement intends to rely on, with the assessments and evaluation drafted for it. The record is unaccepted; accepting reliance on it stays with a person. Returns the record.

      "Record the property valuer report we intend to rely on, with our draft evaluation."

    • atlas_delivery_engagements_independence_declarations_listdelivery:read

      List the independence declarations on an engagement: who declared, the declaration type, the restricted entity list version, any breach identified with its remedial action, and who confirmed it.

      "Has everyone on the audit team made an independence declaration?"

    • atlas_delivery_engagements_independence_declarations_recorddelivery:write

      Record an independence declaration a named person has made on an engagement, including any breach they identified and the remedial action. Record only what the person declared; the caller is recorded as the one who confirmed it. Returns the declaration.

      "Record Priya's engagement-specific independence declaration: no breach identified."

    • atlas_delivery_engagements_rotation_rules_listdelivery:read

      List the partner rotation rules Atlas knows, by jurisdiction and role: the most years a person may serve and the cooling-off years that follow. Use the rule id when recording service on the rotation register.

      "What partner rotation rules apply in the UK?"

    • atlas_delivery_engagements_rotation_register_listdelivery:read

      Read the partner rotation register for one client: each person, their role, the rule that applies, the years served and when they must rotate off. Use this to answer "is the audit partner due to rotate".

      "Is the audit partner for Acme due to rotate?"

    • atlas_delivery_engagements_rotation_register_recorddelivery:write

      Record the years a person has served a client in a role on the partner rotation register, under a jurisdiction rule. An existing entry for the same person, client and role is replaced. Returns the rotation status that follows.

      "Record that Sam has served Acme as engagement partner since 2022."

    • atlas_delivery_engagements_quality_review_getdelivery:read

      Read the engagement quality review of an engagement: why it is required, the reviewer, who appointed them, the unresolved matters and whether it is complete.

      "Who is the engagement quality reviewer on Northwind, and is the review complete?"

    • atlas_delivery_engagements_differences_of_opinion_listdelivery:read

      List the differences of opinion on an engagement, resolved ones included: the nature of each, the escalation path used, the resolution and any dissent recorded.

      "Are there any open differences of opinion on the Acme audit?"

    • atlas_delivery_engagements_differences_of_opinion_recorddelivery:write

      Record a difference of opinion on an engagement, with its nature and the escalation path used. An open difference holds the report date until a person resolves it. Returns the record.

      "Record a difference of opinion on the revenue recognition judgment, escalated to the technical panel."

    • atlas_delivery_engagements_report_dating_readiness_getdelivery:read

      Answer "can the report on this engagement be dated yet". Returns every reason it cannot at once, such as an incomplete quality review or an open difference of opinion. An empty list means nothing is holding it.

      "Can we date the audit report for Acme yet?"

    • atlas_delivery_engagements_non_audit_services_listdelivery:read

      List the non-audit services provided to an audit client on an engagement, with their fees, period and the gaps in each entry, such as a missing pre-approval reference or permissibility rationale.

      "Which non-audit services have we provided to Acme this year?"

    • atlas_delivery_engagements_non_audit_services_recorddelivery:write

      Record a non-audit service provided to an audit client, with its fee and period. The permissibility rationale and the pre-approval reference are left for the partner to record in Atlas, so the entry shows those gaps until they do. Returns the entry with its warnings.

      "Record the tax compliance service we provide to Acme for FY2026, fee 18,000 GBP."

    • atlas_delivery_engagements_tax_profile_getdelivery:read

      Read the tax position of an engagement: the entities and tax residences, place of supply, VAT treatment, e-invoicing mandate, withholding tax and treaty rates, gross-up, certificates required and transfer pricing arrangement.

      "What is the withholding tax position on the Northwind engagement?"

    • atlas_delivery_engagements_tax_profile_setdelivery:write

      Save the tax position of an engagement. Only the fields sent change; the rest keep their values. Rates are decimal strings. Returns the tax profile.

      "Set the treaty withholding rate on Northwind to 5 percent and note a gross-up clause."

    • atlas_delivery_engagements_pe_watches_listdelivery:read

      List the permanent establishment watches on an engagement: each jurisdiction, the day threshold, the alert point, the days counted from time recorded so far, and the risks and mitigations noted.

      "How many days has the team spent in Germany on Northwind against the permanent establishment threshold?"

    • atlas_delivery_engagements_pe_watches_setdelivery:write

      Save the permanent establishment watch for one jurisdiction on an engagement: the day threshold, the day at which to alert, which must come before it, and the risks and mitigations. An existing watch for the jurisdiction is replaced. Returns the watch.

      "Watch Germany for permanent establishment: threshold 183 days, alert at 150."


    Environment variables

    Every option the MCP server reads at boot. Pass them in the host config env block or export them in your shell.

    VariableDefaultPurpose
    ATLAS_API_KEYrequiredPersonal Access Token (atlas_pat_...).
    ATLAS_API_URLhttps://api-atlas.wrxstack.comAtlas API base URL. Override it to point at a staging server.
    ATLAS_MCP_LOG_LEVELinfoStructured-logger threshold. One of debug, info, warn, error.
    ATLAS_MCP_REQUEST_TIMEOUT_MS30000Per-request HTTP timeout against the Atlas API. Bump for slow networks; the agent surfaces a typed timeout error.
    PORT8765 (SSE only)Bind port for the hosted SSE transport.

    Discovery endpoint

    The Atlas API publishes a static descriptor so MCP-aware clients can auto-discover the canonical npm package, its version, and the full tool catalog. This page is rendered from the same descriptor, so it is the single source of truth.

    bash
    curl https://api-atlas.wrxstack.com/.well-known/atlas-mcp.json

    Want the full REST surface? See the OpenAPI 3.1 spec or the Redoc reference.


    Security model

    Every call is authenticated, scope-gated, rate-limited, and audit-logged. Normal-user MCP uses PAT-scoped tools only. Platform-operator MCP is documented separately for authorized Super Admins and is never configured from this page.

    • PAT-only auth. The MCP server has no ambient credentials; it forwards the token from ATLAS_API_KEY on every call. Revoke a PAT and the agent loses access immediately.
    • Scope enforcement. Each tool maps to one required scope (listed above). A PAT without that scope fails with 403 insufficient_scope and the missing scope is named in the body.
    • Idempotency. Write tools (create_task, create_project) send an Idempotency-Key so a retried tool call does not create duplicates within 24h.
    • Audit log. Every PAT mutation is recorded under Activity with the actor, the tool name, and the affected resource.
    • Hosted SSE auth. The SSE transport requires `Authorization: Bearer <ATLAS_MCP_SERVER_AUTH_TOKEN>`; Atlas API calls still use the env-supplied PAT. Platform-operator MCP uses an isolated transport, credential, and route outside this normal-user surface; that bearer must differ from ATLAS_MCP_SERVER_AUTH_TOKEN, and startup fails when they match. Keep bearer tokens separate from PATs and run hosted SSE behind your normal private-network controls.

    Troubleshooting

    The host says "atlas" is connected but no tools appear.
    Restart the host process completely (most hosts only re-scan MCP servers on cold start). On macOS Claude Desktop, that means quitting from the menu bar (Cmd+Q); the dock icon close-window button is not enough.
    Claude Desktop on macOS shows "spawn npx ENOENT" in the logs.
    The macOS GUI does not inherit your shell PATH, so "npx" is not resolvable. Either install Node via the official .pkg installer (which adds /usr/local/bin to the system PATH) or use the absolute path: command: "/opt/homebrew/bin/npx" (adjust to wherever which npx prints in your shell).
    Every call returns "401 Unauthorized".
    The token is missing, expired, or revoked. Mint a fresh one at Settings, API access, update ATLAS_API_KEY, and restart the host.
    Reads succeed but writes fail with 403 insufficient_scope.
    The PAT was minted with read-only scopes. Mint a new token with tasks:write and projects:write added.
    Updates fail intermittently with 409 Conflict.
    The task changed since the agent last read it. The agent should re-read the task (it will get the new version), then retry the update with the fresh ifMatchVersion. Most hosts handle this automatically.
    I am self-hosting Atlas. How do I point the MCP server at my deployment?
    Set ATLAS_API_URL to your API base URL (e.g. https://atlas.acme.internal) in the env block of the host config. The MCP server uses that for every REST call.

    On this page

    • What is MCP?
    • Core concepts
    • Two ways to connect
    • Connect by URL
    • Run it on your machine
    • What each tool tells your host
    • Resources
    • Prompt templates
    • Tools (2326)
    • Environment variables
    • Discovery endpoint
    • Security model
    • Troubleshooting