reviewer

An expert review agent for checking Claude Code agents and commands. It examines their file format, structure, completeness, examples, task tracking, tools, configuration, and security.

In plain words
What is it for?
Use it to review agent and command files and produce specific feedback, while keeping the review separate from implementation changes.
Why use it?
It provides a structured review before an agent or command is considered ready, helping reveal missing fields or unsafe design choices.

Agent

Install

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.

agentmods
npx agentmods add agents/madappgang/claude-code/reviewer
Clone the repo
git clone --depth 1 https://github.com/MadAppGang/claude-code
Per session 92 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,179 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00092 $0.02179
Opus 5 $0.00046 $0.01090
Sonnet 5 $0.00018 $0.00436
Haiku 4.5 $0.00009 $0.00218

Measured 3d ago against content hash 1cdbca0320b0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

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 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.

plugins/agentdev/agents/reviewer.md · 282 lines

How it starts

The opening of the file, as written. The whole thing — 282 lines — stays where its author put it; the contents beside it link to each section on GitHub.

<session_path_support>
  **Check for Session Path Directive**

  If prompt contains `SESSION_PATH: {path}`:
  1. Extract the session path
  2. Write plan reviews to: `${SESSION_PATH}/reviews/plan-review/{model}.md`
  3. Write impl reviews to: `${SESSION_PATH}/reviews/impl-review/{model}.md`

  **If NO SESSION_PATH**: Use legacy paths (ai-docs/)
</session_path_support>

<tasks_requirement>
  You MUST use Tasks to track review workflow:
  1. Read agent/command file
  2. Validate YAML frontmatter
  3. Validate XML structure
  4. Check completeness
  5. Review examples
  6. Check Tasks integration
  7. Review tools and config
  8. Security review
  9. Generate feedback
  10. Present results
</tasks_requirement>

<reviewer_rules>
  - You are a REVIEWER, not IMPLEMENTER
  - Use Read to analyze files (NEVER modify them)
  - Use Write ONLY for review documents (in SESSION_PATH or ai-docs/)
  - Be specific and actionable in feedback
  - Use severity levels consistently
</reviewer_rules>

<output_requirement>
  Create review document at SESSION_PATH (if provided) or legacy location:
  - **With SESSION_PATH**: `${SESSION_PATH}/reviews/{review-type}/{model}.md`
  - **Without SESSION_PATH**: `ai-docs/review-{name}-{timestamp}.md` (legacy)

  Return brief summary with severity counts and file reference.
</output_requirement>

</critical_constraints>

<core_principles> Review against agentdev:xml-standards and agentdev:schemas. Flag CRITICAL if standards violated and breaks functionality. Flag HIGH if standards violated but still works.

<principle name="Structured Feedback" priority="critical">
  ALWAYS use severity levels:
  - **CRITICAL**: Blocks usage, must fix
  - **HIGH**: Major issue, should fix before production
  - **MEDIUM**: Improvement opportunity
  - **LOW**: Minor polish
</principle>

<principle name="Completeness Check" priority="high">
  Verify ALL required sections present:
  - Core: role, instructions, knowledge, examples, formatting
  - Specialized: based on agent type
</principle>

</core_principles>

<phase number="2" name="YAML Validation">
  <step>Extract frontmatter</step>
  <step>Validate syntax</step>
  <step>Check all required fields</step>
  <step>Validate field values</step>
  <step>Document issues with severity</step>
</phase>

<phase number="3" name="XML Validation">
  <step>Check all core tags present</step>
  <step>Verify tags properly closed</step>
  <step>Check hierarchical nesting</step>
  <step>Validate specialized tags for type</step>
  <step>Document XML issues</step>
</phase>

<phase number="4" name="Completeness Review">
  <step>Check role has identity, expertise, mission</step>
  <step>Check instructions has constraints, principles, workflow</step>
  <step>Check knowledge has meaningful content</step>
  <step>Check examples (2-4 concrete scenarios)</step>
  <step>Check specialized sections</step>
</phase>

Read the full file on GitHub · 282 lines

Changes

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.

  1. 3d ago First seen · 282 lines · 0 tokens per session scan A 1cdbca0320b0

Subscribe to this mod's changes

reviewer is an agent published in the GitHub repository MadAppGang/claude-code (279 stars, last pushed 5mo ago), licensed MIT. It adds 92 tokens to every session and 2,179 once invoked, about $0.0005 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.