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.
npx skills add xdg/xdg-claude --skill how-to-code-reviewgit clone --depth 1 https://github.com/xdg/xdg-claudeWrote 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/skills/xdg/xdg-claude/how-to-code-review)<a href="https://agentmods.dev/skills/xdg/xdg-claude/how-to-code-review"><img src="https://agentmods.dev/badge/skills/xdg/xdg-claude/how-to-code-review/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/skills/xdg/xdg-claude/how-to-code-review"><img src="https://agentmods.dev/badge/skills/xdg/xdg-claude/how-to-code-review.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 | $0.00045 | $0.00550 |
| Opus 5 | $0.00023 | $0.00275 |
| Sonnet 5 | $0.00009 | $0.00110 |
| Haiku 4.5 | $0.00005 | $0.00055 |
Grade A, and why
how-to-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 10d 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.
Handling code review requests
Delegate to the code-review subagent (Agent tool, subagent_type: code-review) instead of reading diffs and writing the review inline. Review runs in a forked Opus context — diff reading, file expansion, and reasoning stay out of the main thread; only the report comes back.
When to delegate vs. handle inline
Delegate when:
- The user asks for a code review, PR review, or "look critically" at recent changes.
- The user asks "what's wrong with X" or "find issues in Y" about a defined scope.
- A natural review point has been reached (just-finished feature, pre-commit check).
Handle inline (do not delegate) when:
- The user wants a quick sanity check on a single small snippet visible in conversation. Delegating costs more than reading three lines.
- The user is asking how to fix a specific known issue, not asking for issue discovery.
- The user wants a discussion or back-and-forth about a design choice — that is conversation, not review.
Crafting the delegation prompt
The subagent's default scope is the working tree against the merge base with main. Override only when the user names something else.
Examples:
- User: "review my changes" → Delegate with empty prompt. Default scope handles it.
- User: "code review the auth refactor" → "Review the auth refactor in the current working tree."
- User: "review PR #482" → "Review PR #482."
- User: "look at src/api/handlers.go" → "Review src/api/handlers.go."
- User: "review the last 3 commits" → "Review the commit range HEAD~3..HEAD."
- User: "security-focused review" → "Review the current working tree with emphasis on security (auth, input validation, secret handling, injection)."
Anti-patterns
- Reading the diff yourself and writing the review in the main session. The subagent owns this; doing it inline burns the main context on diff content.
- Asking the subagent for full code blocks or diffs in its report. It returns findings keyed to
file:line— the code is in the working tree. - Delegating mid-discussion before the user has actually asked for review. "Should I refactor X?" is not a review request.
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.
- 10d ago First seen · 41 lines · 45 tokens per session scan A 0f6afe7e2277
how-to-code-review is a skill published in the GitHub repository xdg/xdg-claude (20 stars, last pushed 3d ago), licensed Apache-2.0. It adds 45 tokens to every session and 550 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 skills, from other repositories
autoreview
Pre-commit/ship code review: Codex default; optional Claude or Pi.
rework-rate
Measure and interpret PR rework rate — the emerging 5th DORA metric.
omh-code-review
This is a Hermes-native code-review workflow skill.
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
code-reviewer
Code review specialist focused on patterns, bugs, security, and performance.
full-repo-review
Comprehensive four-wave review of all repo source files, producing a prioritized issue backlog.