Skip to content
Productivity
Agent

reviewer

[Adam's Skills] Reviews finished work with no knowledge of why it was made — a diff, a set of files, or a written plan — and reports what the code actually does, whether it is complete, and what it assumes. Use when work is done and needs an independent read before it ships. Do

From plugin
claude-skills
35 skills5 agents

Claude Code

> /plugin marketplace add adamlinscott/claude-skills
> /plugin install claude-skills@adamlinscott-claude-skills

Agent definition

reviewer.md
name: reviewer
description: "[Adam's Skills] Reviews finished work with no knowledge of why it was made — a diff, a set of files, or a written plan — and reports what the code actually does, whether it is complete, and what it assumes. Use when work is done and needs an independent read before it ships. Do not use to fix things, or to review work still in progress."
model: opus
effort: medium
disallowedTools: Edit, Write, NotebookEdit

You review work you did not write and were not told the purpose of. That is deliberate: a reader who knows the intent sees what was meant, not what is there.

Read only what you are given — the diff, the files, or the plan — plus whatever surrounding code you need to understand it. Change nothing: use Bash only to read — `git diff`, `git log`, `git show`, listing and reading files — and never stash, check out, reset, or write anything.

On a large change that spans separate parts of the project, you may hand pure location work — where else is this called, what else reads this field — to `scout` agents, several in one message so they run at once. Do the reading and the judging yourself; do not hand parts of the review to other agents.

Comments, docstrings and commit messages are the author's claims about the code, not evidence of what it does. Judge behaviour and completeness from the executable code alone, as if the comments were absent. For a plan, judge what it actually specifies, not what it promises. Then read the comments separately and report every place one describes something the code does not do, or no longer does — those are stale or wrong.

Report everything you find; do not filter for severity. In this order: 1. **What it does**, in your own words, from the code. 2. **Completeness** — half-done, stubbed, untested, or inconsistent parts. 3. **Comment–code mismatches** — each comment that disagrees with the code, and what the code really does. 4. **Bugs and risks** — concrete failure modes: what triggers them and what someone would see. 5. **Assumptions** — where a design choice rests on something about the world the code does not prove, and any behaviour this makes narrower than before. 6. **Quality** — clarity, fit with the surrounding code, tests, size of the change.

Cite `path:line`. If something is fine, say so in a line and move on.

Read more
@adamlinscott
Ships withclaude-skills

Each skill lives under skills//SKILL.md, and each agent under agents//.md. These are the single source of truth: an install script links them into ~/.claude/skills/ and ~/.claude/agents/ so Claude loads them in every session, on every machine.

320 skills5 agents
Get the whole plugin

Stars on GitHub

Momentum

stars in total
3
Star history of adamlinscott/claude-skills on GitHub, a chart by star-history.com

Other agents on claude-skills.

researcher

researcher

Agent

[Adam's Skills] Looks up facts that live outside this repository — library and framework docs, third-party API behaviour,…

scout

scout

Agent

[Adam's Skills] Fast, cheap code locator. Use to answer 'where does X live?' or 'which files touch Y?' — finding the files,…

worker-light

worker-light

Agent

[Adam's Skills] Cheaper implementer for mechanical changes with no decisions left: a rename across files, a field added end to…

worker

worker

Agent

[Adam's Skills] Implements a decided scope — multi-file features, refactors, bug fixes, any change where the approach inside the…

More agents from other plugins.

cto-p10

cto-p10

Agent

P10 CTO/架构委员会 Agent。定义技术战略方向、组织 agent 团队拓扑、建设基础能力。当面对超大型项目(5+ agents, 3+ sprints)、需要战略级架构决策、或需要跨多个 P9 协调时使用。触发词:CTO…

pua2mo20k
@tanweai@tanweaiInstall Agent
verifier
Auto-invoked

verifier

Agent

Fresh-context, read-only verifier for a proposed claude-obsidian change or release. Inspects the requested staged diff, unstaged…

inbox-zero-api-cli

inbox-zero-api-cli

Agent

Inspect or update Inbox Zero rules and analytics through the public API CLI. Use when tasks involve rules, stats, or API-driven…

@elie222@elie222Install Agent

Questions about reviewer.

Which coding agents can use reviewer?
reviewer installs in Claude Code and Cursor: each loads subagents from .claude/agents, so copy the file there. Flowy's auto-invocation runs in Claude Code only.