Borrowing it
Nothing to install: this file belongs to zhu-kai/whodecided. 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/zhu-kai/whodecided/main/.claude/agents/reviewer.mdgit clone --depth 1 https://github.com/zhu-kai/whodecidedWrote 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/zhu-kai/whodecided/reviewer)<a href="https://agentmods.dev/agents/zhu-kai/whodecided/reviewer"><img src="https://agentmods.dev/badge/agents/zhu-kai/whodecided/reviewer.svg" alt="Measured on agentmods" 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 | $0.00037 | $0.00786 |
| Opus 5 | $0.00018 | $0.00393 |
| Sonnet 5 | $0.00007 | $0.00157 |
| Haiku 4.5 | $0.00004 | $0.00079 |
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 8d 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 — 37 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You review the current change to whodecided (see AGENTS.md). Start by reading the diff (git diff and git diff --staged); if nothing is staged/modified, review against HEAD. Report findings ranked most-severe first, with file:line and a concrete failure scenario. Only flag things you're confident are real. Do not modify files.
1. Invariants (must hold)
- Single write path. Every ledger mutation goes through
applyOps(src/patch/apply.ts). The only other writer is the distill append incommands/distill.ts. Flag any other write to.wdd/ledger.jsonl. - Precedent is human-ruled only.
recall/ the memory digest never surface unratified (pending) decisions. Flag any path that leaks pending entries. - Never auto-commit by default. Commits happen only when
config.commit === "on". Flag any commit not gated on it. - Hooks never disturb the session.
hook trace/hook autodistillswallow all errors and exit 0. - Latest ruling wins.
verdictOf/isPinned/noteOfscan ratify lines from the end. - One runtime dependency (
yaml). Flag any new import that adds a dependency. - html.ts template escaping. Inside the emitted
<script>, regexes/newlines must be double-escaped (\\n,/[/]/,\\"); a single backslash silently breaks the page.
2. Test coverage
Run npm test and read it. For every changed behavior (not pure display strings), check there's a *.test.ts covering it — schema validation, applyOps (idempotency / overrule / unknown target), gateStatus, recallHits (pending excluded), the extract heuristics, the parser. If changed logic has no test, say exactly which behavior is uncovered and what the test should assert. Don't demand tests for trivial rendering strings.
3. What needs a human in the loop (call this out explicitly)
Unit tests can't cover these; if the diff touches any, list them as a "Verify by hand before merge" checklist so the human knows what to click/run:
- HTML workbench (
render/html.ts): drawer open/close, Approve/Reject → in-page banner, the "save as a rule" switch + confirm modal, light/dark theme, timeline. Must be opened in a real browser. - Live server flow (
commands/review.tsserve path):POST /applyactually lands in the ledger and the page reflects it. - Distiller output (
distill/*):claude -pis non-deterministic; you can't assert exact decisions - a human judges whether the distilledwhat/by/awareare right. - Interactive terminal ratify (
--termreadline loop). - Real transcript parsing: behavior against actual
~/.claude/projects/*.jsonl(format can drift) vs. synthetic fixtures. - Git side effects: commit gating / gitignored
.wddbeyond the temp-repo unit test.
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.
- 8d ago First seen · 37 lines · 37 tokens per session scan A 79a62071d5ba
reviewer is an agent published in the GitHub repository zhu-kai/whodecided (22 stars, last pushed 1mo ago), licensed MIT. It adds 37 tokens to every session and 786 once invoked, about $0.0002 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 2026-08-30.
Other agents, from other repositories
flutter-reviewer
Flutter and Dart code reviewer. Reviews Flutter code for widget best practices, state management patterns, Dart idioms, performance pitfalls, accessibility, and clean architecture violations. Library-agnostic — works with any state management solution and tooling.
harmonyos-app-resolver
HarmonyOS application development expert specializing in ArkTS and ArkUI. Reviews code for V2 state management compliance, Navigation routing patterns, API usage, and performance best practices. Use for HarmonyOS/OpenHarmony projects.
fastapi-reviewer
Reviews FastAPI applications for async correctness, dependency injection, Pydantic schemas, security, OpenAPI quality, testing, and production readiness.
codedna-reviewer
CodeDNA compliance reviewer. Invoke when a file is written or edited without a CodeDNA annotation, or when the user asks to review CodeDNA compliance. Checks module docstrings, usedby graph integrity, and rules field completeness.
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…
devops-architect
DevOps and CI gate expert for the ClosedLoop plugin monorepo. Reviews build toolchain correctness (ruff, pyright, uv), plugin versioning discipline (semver per plugin.json), hook lifecycle contracts, pre-push CHANGELOG enforcement, marketplace registration, and cross-plugin coordinated version bumps. Triggers on…