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/raymondchins/agentmap/agents-mdgit clone --depth 1 https://github.com/raymondchins/agentmapWhat 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 | $0.01163 | $0.01163 |
| Opus 5 | $0.00581 | $0.00581 |
| Sonnet 5 | $0.00233 | $0.00233 |
| Haiku 4.5 | $0.00116 | $0.00116 |
Grade A, and why
agentmap 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 2d 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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
agentmap — Project Instructions
Auto-loaded every Claude Code session in this repo.
Universal rules live at
~/.claude/CLAUDE.md. This file = project-specific only.
⚠️ This is a PUBLIC OSS repo, not one of Ray's apps
Published to npm as @raymondchins/agentmap and to the MCP registry as io.github.raymondchins/agentmap. Consequences that override the usual defaults:
- Code changes go through a PR, not a direct push.
mainis not branch-protected, so nothing stops you — the discipline is the point. CI (.github/workflows/ci.yml) runs on both PRs and pushes tomain, so a direct push is still gated, just gated after the fact. Docs-only changes that touch no shipped file may go direct. - A release is a deliberate act.
chore(release): X.Y.Zcommits bump the version; the tag triggers publish. Don't bump the version as a side effect of a feature PR. package.jsonfilescontrols what ships. A new runtime file that isn't listed there is missing from the published tarball even though it works locally. Current list:agentmap.mjs,mcp.mjs,hooks,skills,.claude-plugin,server.json,NOTICE.- Assume strangers read the diff. No Ray-specific paths, no
/Users/raymondchin/..., no internal project names in source or docs.
Read before editing
CONTRIBUTING.md— the real rules live there: scope, the near-zero-deps rule, the freshness invariant, style, PR process. Read it, don't re-derive it from this file.README.md(50KB) — user-facing surface. The## Commandssection is the CLI contract.ROADMAP.md(96KB) +CHANGELOG.md(78KB) — history and intent. Check before proposing something already decided against.EVAL.md— the measured claims. Any number quoted in the README must be reproducible fromeval/.
Layout
agentmap.mjs — the entire CLI (246KB, single file, by design)
mcp.mjs — MCP server wrapper
hooks/ — the hooks themselves
agentmap-nudge.mjs (Claude Code)
agentmap-codex-nudge.mjs · agentmap-gemini-nudge.mjs
agentmap-cursor-nudge.mjs (Cursor beforeShellExecution gate)
post-commit — the auto-refresh hook shipped to consumers
hooks.json · INSTALL.md
skills/ — the skill payloads + the shared installer
agentmap/SKILL.md — the skill itself
cursor-rule.mdc · opencode-agentmap-nudge.js
install.mjs · install-helpers.mjs · guidance.md
.claude-plugin/ — plugin + marketplace manifests
test/ — 126 test files, `node --test`
eval/ · benchmark/ — the reproducible measurements behind README claims
examples/ · assets/ · docs/
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.
- 2d ago First seen · 69 lines · 1,163 tokens per session scan A 40d655b7f815
agentmap AGENTS.md is an instructions file published in the GitHub repository raymondchins/agentmap (47 stars, last pushed 3d ago), licensed MIT. It adds 1,163 tokens to every session, about $0.0058 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
roam-code AGENTS.md
Instructions for Cranot/roam-code, covering agents.md — roam-code development guide, what this project is, documentation hub, where files go (private vs public) and quality discipline (from internal/dogfood/ + agi-in-md).
rekal-cli CLAUDE.md
Instructions for rekal-dev/rekal-cli, covering rekal cli, soul, standing rules, architecture and key directories.
coograph CLAUDE.md
Instructions for paullukic/coograph, covering claude code instructions, pre-flight (run on every session start), communication style, quick reference and branching strategy.
codemap AGENTS.md
Instructions for stainless-code/codemap, a project described as: Cut AI-agent token waste 90% — query a local SQLite structural index of your JS/TS/CSS codebase with SQL in one round-trip instead of 3–5 file reads. Symbols, imports, calls, components, CSS tokens, coverage, markers. CLI, MCP (21 tools), HTTP, GitHub…
coograph brutal-honesty.instructions.md
Global communication style — applies to all workspaces and conversations.
coograph testing.instructions.md
Conventions for test files — loaded only when editing tests.