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/thisaay/archeyes/claude-mdgit clone --depth 1 https://github.com/thisAAY/archeyesWhat 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 | $0.01284 | $0.01284 |
| Opus 5 | $0.00642 | $0.00642 |
| Sonnet 5 | $0.00257 | $0.00257 |
| Haiku 4.5 | $0.00128 | $0.00128 |
Grade A, and why
archeyes CLAUDE.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 — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ArchEyes — project guide
Bidirectional plan diagrams for AI coding agents. The agent renders a plan as an
interactive architecture graph; the developer edits it directly (drag, reconnect,
draw edges, comment with @mentions, delete); the structured edits flow back to the
agent as JSON. Works with any agent that can run a CLI and read stdout.
Layout
schema/— the protocol and the single source of truth. JSON Schema forplan-graph.json+ the feedback envelope; TS types are generated from the schemas (npm run gen:types), never hand-edited. Also referential-integrity and orphaned-ref checks.cli/— thearcheyesCLI.review <graph.json>spawns a detached localhost server that outlives the foreground poller (survives a ~10-min bash timeout);review --resumere-attaches. DNS-rebinding guard (Host/Origin = localhost) + session token. stdout carries only the feedback JSON; exit 0 = feedback.ui/— React Flow canvas, built to a single self-containeddist/ui/index.htmlviavite-plugin-singlefile. Consumes the design tokens from_ds/(copied intoui/src/tokens/).plugins/archeyes/— the agent skill (skills/review/SKILL.md) and the Claude Code plugin manifest..claude-plugin/marketplace.json(repo root) lists it._ds/,proto/— the Claude Design system (source of the visual language) and the original clickable prototype. Design provenance, not shipped to npm.
The published npm package ships only dist/cli/*.js + dist/ui/index.html + docs
(see files in package.json). Source runs as TS directly in dev via Node's
type-stripping; the published artifact is compiled JS (esbuild).
Design system
The visual language lives in the Claude Design project "ArchEyes Design System"
(id eaa862b4-8285-4d1a-9b6b-2710dee72e9e). _ds/tokens/*.css is a vendored copy of
it; ui/src/tokens/ is a copy of that. The flow is one way:
Claude Design ("ArchEyes Design System") → _ds/tokens/*.css → ui/src/tokens/*.css → vite build
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 · 98 lines · 1,284 tokens per session scan A b4fb3d78080e
archeyes CLAUDE.md is an instructions file published in the GitHub repository thisAAY/archeyes (7 stars, last pushed 1mo ago), licensed MIT. It adds 1,284 tokens to every session, about $0.0064 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
codemap AGENTS.md
Instructions for beadnall/codemap, covering codemap, if you are claude, if you are grok build, if you are another agent and the short version.
codemap copilot-instructions.md
Instructions for beadnall/codemap: This repository holds a skill for turning a codebase into an explorable isometric diagram. The full instructions are in SKILL.md — read that file and follow it when asked to map, diagram or visually explain how a system fits together.
dashmotion CLAUDE.md
Instructions for csthink/dashmotion, covering claude.md — dashmotion and repo conventions — non-obvious, has bitten.
zzon-doc CLAUDE.md
Instructions for HOKlNG/zzon-doc, covering zzon-doc — 인수인계 (claude 세션용), 네이밍 (혼동 주의), 구조, 핵심 계약 (깨지 말 것) and 최초 이식 출처 (역사적 참고 — 이제 독립).
kleinanzeigen-reader CLAUDE.md
Instructions for its-me-prash/kleinanzeigen-reader, covering claude.md — kleinanzeigen-reader, what this repo is, auto-load instructions for claude code, quick command reference and fetch a listing.
genesis-architect CLAUDE.md
Claude Code instructions for maioio/genesis-architect, covering genesis architect - claude code instructions, project structure, how this skill works, development workflow and key constraints (enforced in skill.md).