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 skills add thisAAY/archeyes --skill reviewgit clone --depth 1 https://github.com/thisAAY/archeyesWrote 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/skills/thisaay/archeyes/review)<a href="https://agentmods.dev/skills/thisaay/archeyes/review"><img src="https://agentmods.dev/badge/skills/thisaay/archeyes/review/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/skills/thisaay/archeyes/review"><img src="https://agentmods.dev/badge/skills/thisaay/archeyes/review.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 | $0.00094 | $0.02273 |
| Opus 5.5 | $0.00038 | $0.00909 |
| Sonnet 5.5 | $0.00019 | $0.00455 |
| Haiku 4.5 | $0.00009 | $0.00227 |
Grade A, and why
review 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.
ArchEyes — bidirectional plan diagrams
ArchEyes makes the diagram an input device. You render the plan as a graph; the developer manipulates it directly; their edits come back to you as a structured diff with exact node/edge IDs — zero prose disambiguation. You revise and re-render. Loop until they approve.
You are the parser. ArchEyes' code never reads a plan. You author plan-graph.json
from whatever is at hand: a fresh plan-mode plan, an existing markdown tech spec, the
codebase itself, or current + planned together (which powers the before/after diff).
Don't interview the developer first. The graph is the question — render your best guess and let them fix it on the canvas. That's faster and more precise than terminal Q&A. The one thing the canvas can't express as an edit is a framing choice: altitude (system vs service vs class level) and before/after diff vs single-state graph. Pick one, state which you picked in your terminal message, and ask up front only when you're genuinely torn — fixing altitude visually means the developer deleting and merging a dozen nodes by hand.
The loop
- Author the graph. Write
plan-graph.json(schema below). Keep your prose plan too — it's the same structure, rendered as text, and the accessible representation. - Run the review.
npx archeyes review plan-graph.json(orarcheyes review …if installed). It serves a canvas, opens the browser, and BLOCKS until the developer acts. - Read the feedback. The command prints ONE JSON envelope to stdout and exits.
- exit
0→ an envelope was printed (revise|approve|cancel). Act on it. - exit non-zero → no feedback (idle timeout / error). If it says the server is still
up, run
archeyes review --resumeto re-attach. Otherwise start a new round.
- exit
- Interpret + revise (guide below). Rewrite
plan-graph.jsonAND your prose plan. - Re-render. Run
archeyes review plan-graph.jsonagain. Loop.action: "approve"→ the current graph + plan are final. Stop.action: "cancel"→ stop the loop, keep the plan as-is, ask the developer in the terminal how they want to proceed.
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 · 94 tokens per session scan A 3d6b6a9c34f2
review is a skill published in the GitHub repository thisAAY/archeyes (8 stars, last pushed 3mo ago), licensed MIT. It adds 94 tokens to every session and 2,273 once invoked, about $0.0004 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 2026-08-31.
review compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| reviewthisAAY/archeyes | Skill | 8 | 3mo ago | MIT | A | 94 | Claude Code |
| top-designwondelai/skills | Skill | 2.6k | 6d ago | MIT | A | 113 | Claude Code, Codex |
| steve-jobs-design-reviewwondelai/skills | Skill | 2.6k | 6d ago | MIT | A | 191 | Claude Code, Codex |
| web-typographywondelai/skills | Skill | 2.6k | 6d ago | MIT | A | 128 | Claude Code, Codex |
| design-everyday-thingswondelai/skills | Skill | 2.6k | 6d ago | MIT | A | 132 | Claude Code |
| design-sprintwondelai/skills | Skill | 2.6k | 6d ago | MIT | A | 137 | Claude Code |
| hooked-uxwondelai/skills | Skill | 2.6k | 6d ago | MIT | A | 109 | Claude Code |