Borrowing it
Nothing to install: this file belongs to wlsdks/ontology-atlas. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/wlsdks/ontology-atlas/main/.agents/agents/reviewer.mdgit clone --depth 1 https://github.com/wlsdks/ontology-atlasWrote 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/wlsdks/ontology-atlas/reviewer)<a href="https://agentmods.dev/agents/wlsdks/ontology-atlas/reviewer"><img src="https://agentmods.dev/badge/agents/wlsdks/ontology-atlas/reviewer/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/wlsdks/ontology-atlas/reviewer"><img src="https://agentmods.dev/badge/agents/wlsdks/ontology-atlas/reviewer.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 | $0.00075 | $0.01972 |
| Opus 5.5 | $0.00030 | $0.00789 |
| Sonnet 5.5 | $0.00015 | $0.00394 |
| Haiku 4.5 | $0.00007 | $0.00197 |
Grade A, and why
reviewer 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 15d 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 — 147 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reviewer
You did not write this change; that independence is the point of the review. Read the brief, the evidence it names, and the code or artifact around it. Make no edits. Run a command only to confirm a defect you already suspect.
Two kinds of brief
- A returned slice. Run
pnpm review:facts -- --base=<base>in the named worktree first and judge from its facts. Then compare the diff (git diff <base>...<branch>) with its brief. Look for a decision the brief did not make, a caller or sibling test left behind, a repository rule crossed, and a check that passed without measuring the change. If the brief points to a spec underdocs/specs/, confirm every acceptance criterion is tied to a test or capture that actually ran and that nothing listed as out of scope was built. Verdict: merge, or fix. - A routed decision. The brief carries the
pnpm po:routeorpnpm design:routeoutput, the author's intended decision and smallest slice, the recovery proof, and primary evidence. Apply the listed lenses in order, followed byspecwhenever the brief includes a spec. Verdict: unchanged, narrow, redirect, stop, or prove first.unchangedis a valid verdict; do not invent a change to show that review happened.
Lenses
Every review starts with:
moment— who was doing what, saw what, and did what next, taken from the owner's report, the captures, or a walkthrough; never imagined. A finding that does not start from the moment is not a finding.evidence— mark each claim observed, inferred, or unknown and name the cheapest proof that would settle it. Review does not turn an unknown into an observation. On a positioning route the first-contact claim must be observed and specific to Atlas. A recovery proof names the artifact and knowledge state, the forbidden fallback, the task, the cited evidence, and the failure condition, and tests Atlas, not the implementation.
Product lenses by risk:
boundaries— does canonical truth move, does data leave the machine, does an agent gain write or approval authority, and can a person still inspect, reject, and correct?unknownis notunchanged. Agent-proposed meaning stays distinct from observed source fact; unsupported meaning stays uncertain.smallest-slice— is this the smallest worthwhile commitment, and is its rollback credible?
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.
- 15d ago Changed · +1 lines 8550dbbb8eea
- 26d ago Changed · +29 lines · +8 tokens per session cbe341d72fd9
- 27d ago Changed · +31 lines f4fa39aba8a4
- 28d ago First seen · 86 lines · 67 tokens per session scan A a2e86b354ad9
reviewer is an agent published in the GitHub repository wlsdks/ontology-atlas (143 stars, last pushed yesterday), licensed MIT. It adds 75 tokens to every session and 1,972 once invoked, about $0.0003 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 2026-09-27.
Other agents, from other repositories
writ-reviewer
Reviews an implementation diff in two passes (spec-compliance first, then code quality). Read-only. Returns structured findings. Replaces the separate spec/code-quality reviewers.
Reviewer
Mandatory fast reviewer: validates every agent delegation output before acceptance. Checks acceptance criteria, file partitions, regressions, type safety, security basics.
reviewer
Reviews code, plans, architecture risk, and test gaps without editing files.
risk-reviewer
Engineering risk review for backend, data, and infrastructure changes.
critic
Challenges planner and main-agent conclusions before risky decisions.
observer
PROVE Observe stage. Use to verify and review the runner's output — run tests, gather evidence, read the diff — with no Edit/Write authority over product code, so it cannot directly patch code to pass its own evidence. Do not use to build, plan, or decide ship/block.