code-review

A code-review command that examines a selected change using categories suited to the project. It can review the current branch, another branch, or a GitHub pull request.

In plain words
What is it for?
Use it to review local branch changes, compare against a specified branch, or inspect a GitHub pull request.
Why use it?
It turns a broad review into a structured check of the relevant parts of the change and helps identify the correct comparison point.

Command for Claude Code

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 commands/wbern/agent-instructions/code-review
Clone the repo
git clone --depth 1 https://github.com/wbern/agent-instructions

Made for: Claude Code.

Per session 10 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,265 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 $0.00010 $0.02265
Opus 5 $0.00005 $0.01132
Sonnet 5 $0.00002 $0.00453
Haiku 4.5 $0.00001 $0.00227

Measured 2d ago against content hash 2726718719a2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, 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 2d 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.

.claude/commands/code-review.md · 283 lines

How it starts

The opening of the file, as written. The whole thing — 283 lines — stays where its author put it; the contents beside it link to each section on GitHub.

General Guidelines

Output Style

  • Never explicitly mention TDD in code, comments, commits, PRs, or issues
  • Write natural, descriptive code without meta-commentary about the development process
  • The code should speak for itself - TDD is the process, not the product

Beads is available for task tracking. Use mcp__beads__* tools to manage issues (the user interacts via bd commands).

Plan File Restriction

NEVER create, read, or update plan.md files. Claude Code's internal planning files are disabled for this project. Use other methods to track implementation progress (e.g., comments, todo lists, or external tools).

Code Review

Perform a code review using dynamic category detection.

Phase 0: Setup & Categorization

Determine What to Review

Parse the argument to determine the review target:

Input Action
No argument Detect divergence point, confirm scope with user
Branch name Use specified branch as base
PR number (e.g., 123) Fetch PR diff from GitHub
PR URL (e.g., https://github.com/owner/repo/pull/123) Extract PR number and fetch diff

For GitHub PRs:

  1. Try GitHub MCP first: mcp__github__pull_request_read with method: "get_diff"
  2. Fall back to gh CLI: gh pr diff <number>
  3. If neither works, report error and stop

For local branches (no argument or branch name provided):

  1. Get current branch: git rev-parse --abbrev-ref HEAD

  2. Check for uncommitted changes: git status --porcelain

    • If output is non-empty, note that uncommitted changes exist
  3. Detect divergence point (skip if branch name was provided as argument):

    • Get all local branches except current: git branch --format='%(refname:short)'
    • For each branch, find merge-base: git merge-base HEAD <branch>
    • Count commits from merge-base to HEAD: git rev-list --count <merge-base>..HEAD
    • The branch with the fewest commits back (closest merge-base) is the likely parent
    • If no other branches exist, fall back to main, master, or develop if they exist as remote tracking branches

Read the full file on GitHub · 283 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. 2d ago First seen · 283 lines · 10 tokens per session scan A 2726718719a2

Subscribe to this mod's changes

code-review is a command published in the GitHub repository wbern/agent-instructions (168 stars, last pushed 3mo ago), licensed MIT. It adds 10 tokens to every session and 2,265 once invoked, about $0.0001 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.