ado-eod
MCP server that turns your IDE session history and git activity into Azure DevOps ticket updates, drafting comments, hours, and state changes with confirm-before-post. It also provides read-only progress, people, breakdown, and timeline reports for leads.
README
ado-eod
End your day with one sentence, not twenty minutes of copy-paste.
ado-eod updates your Azure DevOps tickets from inside your AI coding IDE. It reads what
you actually did today — from your IDE's own session history and your git repos — drafts
the ticket comment, hours, and state change, shows you the draft, and posts it after you
say yes. It also answers the manager side: "how did the project go this week?", "what
has everyone been working on?" — straight from Azure DevOps.
Works with Claude Code, OpenAI Codex, Cursor, and Antigravity.
Why
At the end of the day, the only accurate record of your work is inside your AI IDE's chat history. Re-typing it into Azure DevOps — task name, comment, estimation, issues faced, test scenarios — is boring enough that everyone delays it. This tool removes the re-typing. You review a draft and say "yes". That's the whole workflow.
Setup (once, ~2 minutes)
You need Node.js 20+. Then:
npx github:Ramana-Balaji/ado-eod setup
It asks one question — paste your Azure DevOps address (the page where your tickets
are, e.g. https://dev.azure.com/contoso/Contoso%20Web) — and figures out the rest.
If you already know your organization name you can skip the question:
npx github:Ramana-Balaji/ado-eod setup --org contoso --project "Contoso Web"
That one command:
- finds which supported IDEs you have installed and wires each of them up,
- opens your browser once to sign in with your normal work account (Microsoft Entra — no tokens to create or paste),
- confirms everything works and tells you what to try.
Restart your IDE afterwards. Done.
No Azure CLI needed. Sign-in is remembered in your OS's secure store (macOS Keychain / Windows credential store / Linux libsecret), so you won't be asked again for months. If your machine can't use the secure store, set
ADO_EOD_PATto a personal access token as a fallback.
Daily use
Type into your IDE's chat:
update my ticket for today
The assistant collects your day's evidence, drafts each ticket's update, and shows you exactly what will be written:
- a Markdown comment (what you did, repos touched, issues hit, test scenarios, what's next)
- hours:
Completed 4.5h → 7.0h · 43% → 67% done— cumulative, never overwritten - a state change when appropriate
Nothing is posted until you say yes. If the evidence can't fill a required section, it asks you instead of inventing.
When your work is complete, say who tested it:
this is done, tested with alex — update the ticket
The comment then ends with a sign-off handoff that @-mentions the tester:
Status: complete — tested with @Alex. @Alex could you confirm this works as expected and the ticket can be closed?
The tool refuses to set a ticket to Closed or Removed — enforced in the server, not just the prompt. The tester confirms and closes — that's the point.
For leads and admins
Ask in plain language:
how did Fabrikam Web go this week? what has Jordan been working on?
Four read-only views: progress (state movement, hours), people (who's doing what, last activity), breakdown (epic → feature → story tree), timeline (dated sequence of changes and comments). You see exactly what Azure DevOps already lets you see — the tool uses your sign-in, so its permissions are your permissions.
Rules (for the admin)
Behaviour is governed by a layered rules file — rules.default.yaml (shipped) →
org file → ~/.ado-eod/rules.yaml (per machine), later files win per key:
| Rule | What it controls |
|---|---|
ado.* |
organization, default project, and the regex that finds ticket ids in branches/commits |
repoRoots |
where to look for git repositories when attributing work |
applies.* |
which projects, work item types, and states may be updated; onlyMyTickets refuses drafts for tickets assigned to someone else |
comment.required |
sections the draft must fill — missing ones are flagged and the assistant asks instead of posting hollow entries |
comment.template / comment.signoffTemplate |
the comment format and the completion handoff |
completion.* |
the maximum state the tool may propose; whether a tester is required |
testScenarioField |
which work item field test scenarios are appended to, per type |
hours.* |
daily cap (default 14h), idle-gap threshold, rounding |
fields.* |
which fields may be written at all (hours, state, long-text) |
redact.extraPatterns |
additional redaction patterns (base patterns can never be removed) |
Two things are hard-coded and no rules file can disable them: confirm-before-post, and the base redaction patterns (passwords, tokens, connection strings and the like never leave your machine).
What it reads, what it writes
Reads (local): your own IDE session logs (~/.claude, ~/.codex, ~/.cursor) and
git log of your repos. File contents are never included — only paths. All mined text
(your prompts, commit subjects) is scrubbed against redaction patterns before it can
appear in a draft — and you review every draft before anything is posted.
Writes (Azure DevOps, only after you confirm): one comment per ticket per day (re-runs replace it — no double-posting, no double-counted hours), the two hour fields, optionally a state change and appends (never overwrites) to Description / Repro Steps / Acceptance Criteria.
The tools (MCP)
| Tool | Access | Purpose |
|---|---|---|
eod_worklog |
local only | the day's evidence bundle |
eod_draft |
reads ADO | per-ticket draft: comment, hours, %, state |
eod_status |
reads ADO | diagnostics — auth, config, rules in force |
eod_report |
reads ADO | progress / people / breakdown / timeline |
eod_post |
writes ADO | posts a confirmed draft; refuses without confirmed: true |
eod_create |
writes ADO | creates a work item; refuses without confirmed: true |
Development
git clone https://github.com/Ramana-Balaji/ado-eod
cd ado-eod
npm install
npm test
No frameworks: TypeScript, node:test, and five small dependencies
(@modelcontextprotocol/sdk, @azure/identity, @azure/identity-cache-persistence,
yaml, zod).
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.