code-review

code-review is a skill for Claude Code from hg-pyun/claude-code-marketplace. It costs 104 tokens per session (2,337 once invoked), scanned A, original, MIT.

A structured code-review skill that examines changes and reports findings by severity.

In plain words
What is it for?
It is for reviewing staged or unstaged changes, or specified files, for correctness, security, performance, code quality, and documentation concerns.
Why use it?
It helps developers focus first on critical problems, then on major and minor issues, instead of receiving an unclear general opinion.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the dev-tools plugin — 11 skills, 1 command, 15 agents shipped together

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 skills/hg-pyun/claude-code-marketplace/code-review
Any agent
npx skills add hg-pyun/claude-code-marketplace --skill code-review
Clone the repo
git clone --depth 1 https://github.com/hg-pyun/claude-code-marketplace

Made for: Claude Code.

Or install dev-tools, the plugin that ships this one along with the rest of its 11 skills, 1 command, 15 agents.

Wrote 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.

agentmods badge for code-review

README.md
[![agentmods](https://agentmods.dev/badge/skills/hg-pyun/claude-code-marketplace/code-review.svg)](https://agentmods.dev/skills/hg-pyun/claude-code-marketplace/code-review)
Your own site
<a href="https://agentmods.dev/skills/hg-pyun/claude-code-marketplace/code-review"><img src="https://agentmods.dev/badge/skills/hg-pyun/claude-code-marketplace/code-review.svg" alt="Measured on agentmods" height="20"></a>
Per session 104 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,337 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 5.1 $0.00104 $0.02337
Opus 5 $0.00052 $0.01169
Sonnet 5 $0.00021 $0.00467
Haiku 4.5 $0.00010 $0.00234

Measured 6d ago against content hash 5d23bbfd52a3, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

code-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 6d 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/dev-tools/skills/code-review/SKILL.md · 111 lines

How it starts

The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.

<Use_When>

  • User asks for code review of staged/unstaged changes
  • User says "review my changes", "/code-review", "코드 리뷰 해줘", "이거 리뷰"
  • User wants severity-rated feedback rather than free-form impression
  • Before opening a PR, when the user wants a self-review pass </Use_When>

<Do_Not_Use_When>

  • User asks a syntax question — answer directly
  • User wants explanation of existing code, not review
  • User wants to apply fixes, not surface findings — use direct edit </Do_Not_Use_When>

<Why_This_Exists> Free-form "looks good" reviews consume time without informing decisions. Severity-rated reviews let the user triage: fix CRITICALs now, batch MAJORs, defer MINORs. This skill is the canonical entrypoint for that triage style in the dev-tools plugin. The reviewer advisor is always dispatched (review stays a separate pass); security-auditor and doc-writer join the panel only when the diff actually touches their surface, so small diffs are not taxed with irrelevant advisor passes. </Why_This_Exists>

<Execution_Policy>

  • Always identify the review surface first (diff scope, or named files).
  • Always dispatch the reviewer agent — review is a separate pass; do NOT inline the review logic.
  • Route the other two advisors by what the diff actually touches, using the git diff --stat output from Step 1 as the gate:
    • security-auditor: dispatch only when the diff touches a security surface (auth/crypto/session code, input handling, serialization, config or secret files) or the diff is large (>150 changed lines).
    • doc-writer: dispatch only when the diff touches *.md files or changes a public API signature, option, or config schema.
    • A config-schema change that is also security-sensitive (auth, secrets, permissions config) trips BOTH gates — dispatch security-auditor and doc-writer together.
    • For every advisor not dispatched, note it in the output as one line: <agent>: skipped (no <surface> in diff).
  • Slug-context exception: when a slug context is active (hand-off persistence path), always dispatch the full three-advisor panel — no routing gate — so persisted artifacts keep consistent multi-domain coverage.
  • When security-auditor is on the panel, append to the reviewer's dispatch prompt: "security-auditor is on this panel — skip the Security checklist category and focus on logic/quality/spec-compliance". (reviewer.md itself is unchanged — a standalone reviewer invocation runs its full checklist.)
  • Output must use the severity bar consistently across invocations.
  • Hand-off persistence (slug context only): when a slug context is active (e.g., autopilot invocation or --slug=<slug> argument), each advisor's findings MUST also be persisted to .dt-handoff/<slug>/artifacts/ask/<agent>-<ISO8601>.md with the descriptor frontmatter (kind=advisor, retention=session, status=complete; schema: scripts/validate.sh). For free-standing user invocations without a slug, in-session text output is sufficient and no file is written. </Execution_Policy>

Read the full file on GitHub · 111 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. 6d ago First seen · 111 lines · 104 tokens per session scan A 5d23bbfd52a3

Subscribe to this mod's changes

code-review is a skill published in the GitHub repository hg-pyun/claude-code-marketplace (2 stars, last pushed 2mo ago), licensed MIT. It adds 104 tokens to every session and 2,337 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 2026-08-31.