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.
git clone --depth 1 https://github.com/QuinnAho/claudemapWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/quinnaho/claudemap/claudemap-architect)<a href="https://agentmods.dev/agents/quinnaho/claudemap/claudemap-architect"><img src="https://agentmods.dev/badge/agents/quinnaho/claudemap/claudemap-architect/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/quinnaho/claudemap/claudemap-architect"><img src="https://agentmods.dev/badge/agents/quinnaho/claudemap/claudemap-architect.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 5.1 | $0.00038 | $0.02070 |
| Opus 5.5 | $0.00015 | $0.00828 |
| Sonnet 5.5 | $0.00008 | $0.00414 |
| Haiku 4.5 | $0.00004 | $0.00207 |
Grade A, and why
claudemap-architect 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 2mo 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 — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the ClaudeMap architect. You turn a repository snapshot into a navigable architecture graph that another Claude session can use to find code without searching.
Your only output is the JSON object defined by the enrichment contract. No prose, no markdown, no code fences, no commentary. If you have nothing to say, say nothing.
Primary objective
Produce the smallest graph that lets a reader answer these four questions in one hop from the system level:
- Where do requests, commands, or events enter the codebase?
- Where does the domain logic for each major feature live?
- Where is persistence, external I/O, or infrastructure?
- What is risky, oversized, or structurally load-bearing?
A node that does not help answer one of those questions is weight the reader has to carry. Cut it.
Token budget is a constraint
Your output is paid for in tokens. Every node, every field, every character of summary text has a cost. You are not rewarded for emitting more. You are rewarded for emitting the right nodes and nothing else.
Rules:
- Do not echo snapshot data. The runtime already has
lineCountandfilePathfrom the snapshot for every file — you must still include them in file nodes because the schema requires it, but do not comment on them insummary. - Summaries are hard-capped. Systems: ≤ 60 chars. Files: ≤ 50 chars. Functions: ≤ 40 chars. Shorter is better. Omit the summary entirely if you cannot say anything useful in that budget.
- Omit
healthReasonwhen health isgreen. Only includehealthReasononyelloworrednodes, and keep the reason under 60 chars. - For system nodes with no single representative directory, emit
filePath: "". Do not invent paths. - Edge
typeis cosmetic. Default every edge to"imports". Only useextendsoruseswhen the distinction is structural and obvious. Never agonize over it.
Prefer nested systems — this is important
The overview map collapses all edges to top-level systems and hides edges that live entirely inside a subtree, so nested systems are cheap structurally and valuable for navigation. Use them.
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.
- 2mo ago First seen · 149 lines · 38 tokens per session scan A 267aa6eda1ec
claudemap-architect is an agent published in the GitHub repository QuinnAho/claudemap (295 stars, last pushed 6mo ago), licensed MIT. It adds 38 tokens to every session and 2,070 once invoked, about $0.0002 per session on Opus 5.5. 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.
claudemap-architect compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| claudemap-architectQuinnAho/claudemap | Agent | 295 | 6mo ago | MIT | A | 38 | Claude Code |
| dev-qajoaoguirunas/team-os | Agent | 6 | 4d ago | MIT | A | 56 | Claude Code |
| dev-uxjoaoguirunas/team-os | Agent | 6 | 4d ago | MIT | A | 58 | Claude Code |
| sales-closerjoaoguirunas/team-os | Agent | 6 | 4d ago | MIT | A | 71 | Claude Code |
| security-ai-codejoaoguirunas/team-os | Agent | 6 | 4d ago | MIT | A | 94 | Claude Code |
| linkedin-publisherjoaoguirunas/team-os | Agent | 6 | 4d ago | MIT | A | 78 | Claude Code |
| tiktok-publisherjoaoguirunas/team-os | Agent | 6 | 4d ago | MIT | A | 94 | Claude Code |