Skip to content
AI & Agents
Skill

/review

Closeout pass on a finished substantive change — sequence refactor-clean (shape), code-review (diff), and write-docs (docs) so all three run, in the order where each feeds the next. Use after implementing a slice or feature and before calling it done or committing; when the user

BOOST
From plugin
dzhng-skills
1.3k30 skills
Install
$ npx -y skills add dzhng/skills --skill review --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/review

Context preview

The summary Claude sees to decide when to auto-load this skill.

Closeout pass on a finished substantive change — sequence refactor-clean (shape), code-review (diff), and write-docs (docs) so all three run, in the order where each feeds the next. Use after implementing a slice or feature and before calling it done or committing; when the user

SKILL.md

review.SKILL.md
name: review
description: Closeout pass on a finished substantive change — sequence refactor-clean (shape), code-review (diff), and write-docs (docs) so all three run, in the order where each feeds the next. Use after implementing a slice or feature and before calling it done or committing; when the user asks for a final review, a cleanup-and-document pass, or to "review everything." For a diff-only audit, reach code-review directly.

Review

A finished change earns "done" only after three lenses pass: its **shape**, its **diff**, and its **docs**. Each has a specialist skill. This skill owns only what they can't — the order they run in, the loop between them, and one verdict.

Workflow

1. **Scope.** Fix the diff under review and confirm it's substantive; skip the whole pass for trivial edits. 2. **Shape — [refactor-clean](../refactor-clean/SKILL.md).** Run it first, while restructuring is cheap and before you audit a diff that's about to move. 3. **Diff — [code-review](../code-review/SKILL.md).** Audit the settled shape. A finding that forces a structural change sends you back to step 2, not onward. 4. **Docs — [write-docs](../write-docs/SKILL.md).** Update the docs this change touched. Then trace the link chain from the root README down to each one: every link on the path still resolves, and the hub-to-leaf flow still reads in order. 5. **Report.** One verdict across all three passes — what each found, what you changed, what you left and why. Done only when every pass is clean or resolved.

Rules

  • When you enter a pass, read its skill — the rules live there, not here.
  • Order governs presentation, not disclosure: surface a finding the moment you

hit it, even when a later pass owns it.

  • Loop, don't cascade: a later fix that reopens an earlier pass returns there.

Stop when a full pass adds nothing.

Read more
Ships withdzhng-skills

AI skills for building software factories. My personal library of domain-agnostic agent skills, reused across every project.

Get the whole plugin
Stats
1,337
Stars
81
Forks
Active
Maintenance
HTML
Language
MIT
License
13h ago
Last commit
3mo ago
Created
5h ago
Added

Repo: dzhng/skills

Other skills on dzhng-skills.