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/beadnall/codemap/agents-mdgit clone --depth 1 https://github.com/beadnall/codemapWhat 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 | $0.00747 | $0.00747 |
| Opus 5 | $0.00374 | $0.00374 |
| Sonnet 5 | $0.00149 | $0.00149 |
| Haiku 4.5 | $0.00075 | $0.00075 |
Grade A, and why
codemap 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 — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
codemap
Turn a codebase into an explorable diagram: isometric blocks on a grid, a component index, a what-it-does / how-it's-built panel, and the system's own data moving along the edges as dots the reader can stop and inspect.
The instructions are in SKILL.md. Read that file, then follow it. This page exists so that agents which do not read Anthropic's skill format can still find them.
If you are Claude
You already have this as a skill if it is installed in ~/.claude/skills/ or
<repo>/.claude/skills/. Nothing here to do.
If you are Grok Build
Install the skill into one of the discovery paths:
~/.grok/skills/codemap/ # or <repo>/.grok/skills/codemap/
Then ask for a map normally — "map this codebase", "codemap this repo",
"diagram the architecture". Grok Build picks up SKILL.md automatically, so
there is no separate instructions file to copy.
Delivery. Write the finished HTML to disk, for example codemap.html, and
tell the user the path. The page is self-contained and opens in any browser
with no server.
If you are another agent
Nothing in the method is Claude-specific. The template is plain HTML and JavaScript, the scripts are Python, and the instructions are markdown. Two details differ:
- Triggering. There is no frontmatter for you to match on, so this only runs when the user asks for it and points you here.
- Delivery. SKILL.md's last step suggests an artifact or canvas if you have one. If you do not, write the file to disk and tell the user the path. It is self-contained by design and opens in any browser with no server.
Everything else — mining the repository for facts, choosing nodes and edges, filling the four arrays, verifying, rendering — applies unchanged.
The short version
cp assets/template.html my-map.html
# replace STATS, NODES, EDGES and REGIONS, plus <title> and the <svg aria-label>
python3 scripts/verify.py my-map.html
python3 scripts/render.py my-map.html render.png
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 · 79 lines · 747 tokens per session scan A af5dbebaae58
codemap AGENTS.md is an instructions file published in the GitHub repository beadnall/codemap (8 stars, last pushed 17d ago), licensed MIT. It adds 747 tokens to every session, about $0.0037 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 2026-08-31.
Other instructions, from other repositories
travel-plan-viz CLAUDE.md
Instructions for zexuanw958-svg/travel-plan-viz, covering claude.md, 这是什么, 架构红线, 测试 and 数据采集约束（写在 references/research-guide.md，改动要同步）.
attention-control AGENTS.md
Instructions for aaddrick/attention-control, covering working in this repository, the one rule that breaks the build, which file your tool reads, write in the style you are editing and verify every claim about a harness.
attention-control GEMINI.md
Instructions for aaddrick/attention-control: Shape every response for a reader with ADHD, then write each sentence in controlled English. Follow the rules imported below in full: lead with the next action, number multi-step work, restate state every turn, suppress tangents, cut all preamble and closers, and keep one…
talisman-skill AGENTS.md
Instructions for jeunecraack/talisman-skill, covering talisman — agent playbook (v1), when to run, workflow (8 phases), key features and repo structure.
dashmotion CLAUDE.md
Instructions for csthink/dashmotion, covering claude.md — dashmotion and repo conventions — non-obvious, has bitten.
cinematic-scroll-skill AGENTS.md
Instructions for MustBeSimo/cinematic-scroll-skill, covering agents.md, what it is, when to use it, canonical contract and trigger usage.