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.
npx agentmods add instructions/composiohq/composio/agents-mdgit clone --depth 1 https://github.com/ComposioHQ/composioWhat 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 | $0.01143 | $0.01143 |
| Opus 5 | $0.00571 | $0.00571 |
| Sonnet 5 | $0.00229 | $0.00229 |
| Haiku 4.5 | $0.00114 | $0.00114 |
Grade A, and why
composio 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 3d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Repository guidance for Codex, Claude Code, Cursor, pi.dev, and other AI coding agents.
Scope
This is the Composio SDK v3 monorepo. Most product code lives in ts/ and python/; docs live in docs/. Branch new work from next and target PRs at next unless the user says otherwise.
First Steps
- Read the nearest nested
AGENTS.mdbefore editing a subtree. - Preserve unrelated dirty work. Do not revert, delete, or reformat files outside the requested scope.
- Treat
.agents/skillsas the canonical local skill tree..claude/skillsis a compatibility symlink and must not be edited as a separate copy. - Do not edit generated SDK surfaces or vendor trees — see Generated And Vendored Paths.
- Verify every command you write against the current
package.json,Makefile,noxfile.py, or workflow file.
Skill Routing
Use the smallest relevant skill:
repo-guidance: repo layout, branch/PR workflow, changesets, generated files, or cross-repo maintenance.skill-maintenance: adding or changing local skills, skill references, compatibility mirrors, or validation.bug-fixing: fixing a defect and choosing regression tests.cross-sdk-parity: aligning TypeScript and Python behavior, generated client bumps, or API contract drift.docs-decisions: docs site work, changelogs, docs decisions, or docs automation.good-docs-writing: drafting or revising docs prose in the house writing voice.good-docs-audit: reviewing existing docs or prose against that voice and reporting violations.eve: durable backend AI agents built with the eve framework.typescript-sdk,typescript-testing,typescript-providers: TypeScript SDK/core/provider work.cli-command,cli-e2e: CLI design/implementation or CLI end-to-end tests.cli-release: first-party CLI beta builds, stable promotion, release verification, or recovery.python-sdk,python-testing,python-providers,python-release: Python SDK/provider/testing/release work.
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.
- 3d ago First seen · 106 lines · 1,143 tokens per session scan A ed82ccaff5a3
composio AGENTS.md is an instructions file published in the GitHub repository ComposioHQ/composio (30,003 stars, last pushed yesterday), licensed MIT. It adds 1,143 tokens to every session, about $0.0057 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.
Other instructions, from other repositories
PraisonAI AGENTS.md
AGENTS.md instructions for MervinPraison/PraisonAI, covering agent instructions, project guidelines and issue and pr routing.
asm AGENTS.md
Instructions for luongnv89/asm, covering agents.md, when to delegate, definitions, shared boundaries and editing this file.
asm CLAUDE.md
Instructions for luongnv89/asm, covering claude.md, critical commands, architecture map, hard rules and workflow preferences.
noclick CLAUDE.md
Claude Code instructions for noclickapp/noclick, covering claude.md, what this is, layout, commands and the one thing to understand first: schemas are generated.
claude-nexus-hyper-agent-team-light CLAUDE.md
Instructions for asiflow/claude-nexus-hyper-agent-team-light, covering claude.md, the cto — top authority, gold prompt (maximum team activation), comprehensive dispatch table and dispatch format.
noclick AGENTS.md
AGENTS.md instructions for noclickapp/noclick, a project described as: Build agents to automate any background task. Works with your ChatGPT/Claude subscription.