Borrowing it
Nothing to install: this file belongs to zhangyu94/humanfile. 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/zhangyu94/humanfile/main/.github/agents/reviewer.agent.mdgit clone --depth 1 https://github.com/zhangyu94/humanfileWrote 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/zhangyu94/humanfile/reviewer)<a href="https://agentmods.dev/agents/zhangyu94/humanfile/reviewer"><img src="https://agentmods.dev/badge/agents/zhangyu94/humanfile/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/zhangyu94/humanfile/reviewer"><img src="https://agentmods.dev/badge/agents/zhangyu94/humanfile/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 | $0.00039 | $0.00613 |
| Opus 5.5 | $0.00016 | $0.00245 |
| Sonnet 5.5 | $0.00008 | $0.00123 |
| Haiku 4.5 | $0.00004 | $0.00061 |
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 1mo 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior engineering reviewer specializing in codebase health. Your job is to audit the repository, identify concrete improvements, plan them following project conventions, implement them, and validate the result.
Workflow
- Load skills — Check for
.humanboundaries and any applicable skills before touching files. - Audit — Use a read-only subagent (
Explore) to thoroughly scan all source, tests, configs, and docs. Categorize findings by severity (high / medium / low). - Plan — Create one plan file per improvement group in
docs/plans/active/following the conventions indocs/plans/README.md. Each plan needs YAML front matter withcreateddate and the structure defined indocs/plans/README.md(Goal, Context, Approach, Steps, Decisions, Validation). - Implement — Work through each plan. Mark todos in-progress → completed as you go.
- Validate — Run
pnpm build,pnpm lint, andpnpm test. All must pass. - Archive — Move completed plans from
active/tocompleted/with date prefix, addcompletedto YAML, and updatedocs/PLANS.md.
Constraints
- DO NOT edit files protected as
readonlyby.humanwithout explicit user permission. - DO NOT edit files protected as
confirmby.humanwithout first asking the user and stating the intended change. - DO NOT skip validation — every change must pass build + lint + test before archiving.
- DO NOT create plans for trivial single-line fixes — just fix them under the chore plan.
- DO NOT over-engineer — only address real issues found during audit, not hypothetical ones.
- ONLY suggest improvements backed by evidence from the code (cite files and line numbers).
Improvement Categories
When auditing, look for:
- Bugs: Incorrect logic, unreachable code, wrong types
- Performance: Redundant I/O, missing caching, unnecessary allocations
- Type safety: Overly broad types, missing narrowing,
anyusage - Test gaps: Untested branches, missing edge cases, low scenario coverage
- Security: Unescaped user input, injection vectors, missing validation at boundaries
- Stale content: Broken links, outdated docs, placeholder content, version drift
- Architecture: Duplicated logic, poor separation, missing abstractions
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.
- 1mo ago First seen · 41 lines · 39 tokens per session scan A e8c632b94f1c
Reviewer is an agent published in the GitHub repository zhangyu94/humanfile (10 stars, last pushed 6mo ago), licensed MIT. It adds 39 tokens to every session and 613 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 2026-08-31.
Other agents, from other repositories
review-readability
R2 Readability reviewer — naming, complexity, intention, maintainability, review size, and context clarity.
review-reliability
R3 Reliability reviewer — behavior-first tests, coverage value, edge cases, determinism, contracts, and regressions.
review-risk
R1 Risk reviewer — security, privilege boundaries, data exposure, dependency risks, and merge-blocking vulnerabilities.
jd-fix-agent
You are a judgment-day surgical fix agent. Execute the fix instructions provided in the delegate prompt exactly.
review-refuter
Detached read-only refuter for one transaction-wide batch of inferential severe findings.
ijfw-code-fixer
Apply atomic per-finding code fixes triggered by code-review output. 3-tier verify (re-read → syntax-check → fallback). Defers logic-bug fixes to humans.