What is mcp AGENTS.md?

Repository instructions for AI coding agents working on a public MCP project. They describe the project, setup and run commands, and what information may be included in public changes.

Use them when changing, testing, documenting or reviewing this repository, especially before committing publicly visible work.

They reduce mistakes by making project rules and public-information limits explicit. This helps agents avoid exposing private services or internal implementation details in code and documentation.

mcp AGENTS.md — Instructions file by Teamwork

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add instructions/teamwork/mcp/agents-md

Clone the repo

git clone --depth 1 https://github.com/Teamwork/mcp

Made for: Codex, OpenCode.

Per session 16,433 This file is loaded in full into every session.

When invoked 16,433 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.16433	$0.16433
Opus 5	$0.08217	$0.08217
Sonnet 5	$0.03287	$0.03287
Haiku 4.5	$0.01643	$0.01643

Measured yesterday against content hash 2848e29caef1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 261 lines

How it starts

The opening of the file, as written. The whole thing — 261 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

A concise guide for AI coding agents working on this repository. It complements README.md with machine-actionable build, test, run, and convention notes.

This repository is public

Everything committed here is world-readable and permanent — source, comments, tests, docs, commit messages, PR titles and descriptions, issue text. A later fix does not retract what has already been cloned, cached or indexed. This file is public too, and is bound by its own rules.

Using an internal source to get the behaviour right is fine; publishing it is not. Only the externally observable result belongs in the repo.

Never name private repositories or services. desksdkgo, twapi-go-sdk and the MCP SDK are public and may be cited freely. The backends behind the Projects and Desk APIs are not — do not name their repositories, packages, files, handlers, models, or query-argument identifiers. Describe the HTTP endpoint instead: route, parameters, JSON keys, status codes.
Never name internal infrastructure. Deployments, hostnames, dashboards, monitoring, CI tooling, chat channels, wikis, runbooks, or environment detail beyond the documented TW_MCP_ENV values.
Never reference internal tracking. Ticket keys, incident IDs, or links and commit hashes in private repositories. Public protocol references (SEP-2575) are fine — those are published.
Never describe a backend defect as a defect. State what the endpoint does today ("list endpoints clamp pageSize over 100 to 100"), not that it is a bug, what causes it, or when it is due to be fixed — the service is still running the old code.
Never publish unreleased information. Roadmap items, planned deprecations, dated fixes, feature flags, or capacity and traffic figures.
Never commit real data. Tokens, API keys, cookies, real installation subdomains, or real customer and employee names, emails and IDs — in code, tests, fixtures, log samples and error strings alike. Fixtures use example.com, test.teamwork.com, [email protected], and IDs like 777/12345.

Read the full file on GitHub · 261 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

yesterday First seen · 261 lines · 16,433 tokens per session scan A 2848e29caef1

Subscribe to this mod's changes

mcp AGENTS.md is an instructions file published in the GitHub repository Teamwork/mcp (25 stars, last pushed 3d ago), licensed MIT. It adds 16,433 tokens to every session, about $0.0822 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Testing ai artificial-intelligence mcp mcp-server mcp-tools teamwork

mcp AGENTS.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

AGENTS.md

This repository is public

What this file has done since we first saw it

Other instructions, from other repositories

spiceai copilot-instructions.md

ai AGENTS.md

ai CLAUDE.md

albucore AGENTS.md

OwnPilot AGENTS.md

agent-pulse AGENTS.md