R-Hatim/hrih-skills / skills/review-apprentice
review-apprentice
VerifiedTurn code review into teaching when an AI reviews alongside people, so juniors learn to review instead of rubber-stamping the agent. Use it when a junior or new team member reviews a pull request with an AI reviewer, when a senior wants to mentor through reviews, or for requests like 'review this PR with me', 'help me learn to review', 'check my review', 'what did I miss?', 'train me on this codebase', and when a solo developer wants to keep their own reviewing sharp. The junior reviews first without seeing the AI's findings, then accepts or rejects each AI finding with a reason, then a senior (or, solo, the lessons file) checks those calls. Every correction becomes a plain-language lesson in REVIEW-LESSONS.md that the next reviewer, human or agent, reads. Even when asked to just give the answers, it offers hints first and the list on request. Skip it when the goal is fast triage of many changes (use review-router) or nobody wants to learn from the review.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.8 KB |
| references/lessons.md | markdown | 7.1 KB |
| references/review-questions.md | markdown | 2.3 KB |
| scripts/diff_facts.py | script | 19.0 KB |
| scripts/lessons.py | script | 18.9 KB |
references/review-questions.md
View on GitHub# Questions a good reviewer asks
A checklist for the blind review. It asks questions and gives no answers, because finding the answer yourself is the part that sticks. You won't need all of them for every change: start with the first section, then pick the ones that fit.
## Before reading the code
- In one sentence, what is this change supposed to do? If you can't say, ask the author before reviewing.
- Who notices if it's wrong: a user, another team, a nightly job, nobody until the end of the month?
- If it breaks in production, how would anyone find out? How would they undo it?
## Does it do what it says?
- Does the diff match the description? Is there anything in it the description doesn't mention?
- Walk one real input through the new code, line by line. Where does it go?
- What happens with the empty case, the one-item case, and the very large case?
- What happens on the second run? Is it safe to retry?
## Where it can go wrong
- Every place that can fail (network, disk, parse, permission): what happens next? Is the error handled, passed on, or swallowed?
- Does anything wait forever, retry forever or grow forever?
- Is any input trusted that comes from outside: a user, a file, a request, another service?
- Are permissions checked where the data is read, or only where the button is?
- Are times, time zones, currencies or units mixed anywhere?
- Can two of these run at the same time? What do they share?
## The tests
- Which test would fail if the main line of this change were deleted? If none, what is being tested?
- Do the tests check the result, or only that the code ran?
- Did any existing test change? Was it wrong before, or was it bent to fit the new code?
- What did the author test by hand that isn't in the tests?
## Fitting in
- Does this do something the codebase already does elsewhere, differently?
- Will the next person to read this understand it without the PR description?
- Is there a house rule (a lesson) that applies here? Does the change follow it?
## Writing it up
- Lead each finding with what goes wrong and for whom, not with what you'd have done instead.
- Mark what's blocking (important), what's optional (nit), and what's a question.
- If you're unsure, ask: "What happens if `owner` is null here?" is a better comment than a guess.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |