CherryHQ/cherry-studio / .agents/skills/gh-pr-review
gh-pr-review
UnverifiedAutomated Cherry Studio review for local branches, PRs, commits, files, architecture docs, and repository skills. Use for code or documentation reviews that need project-specific naming, main/renderer/shared placement and dependency rules, IpcApi and DataApi boundaries, lifecycle/service ownership, renderer hooks, React/UI conventions, and tests. Review depth adapts to diff size and runtime subagent capability (single-agent or multi-agent reviewer-verifier). Report-only by default; code fixes and GitHub submission each require explicit invocation-time authorization (`fix` / `submit`). Normal-review prompts and safe interruption behavior follow the interaction contract below. To diagnose gaps in the skill after a review session, run `/gh-pr-review diag`.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 15.4 KB |
| agents/openai.yaml | config | 247 B |
| references/checklist-evolution.md | markdown | 4.7 KB |
| references/cherry-review-guidance.md | markdown | 40.0 KB |
| references/code-checklist.md | markdown | 10.3 KB |
| references/consumer-review.md | markdown | 7.4 KB |
| references/diagnosis.md | markdown | 2.6 KB |
| references/doc-checklist.md | markdown | 3.8 KB |
| references/judgment-matrix.md | markdown | 6.8 KB |
| references/local-review.md | markdown | 6.2 KB |
| references/pr-review.md | markdown | 19.7 KB |
| references/teams-review.md | markdown | 19.0 KB |
references/checklist-evolution.md
View on GitHub# Checklist Evolution
Rules for evolving review checklists without confusing a review-session
suggestion with durable repository knowledge. Goal: keep checklists minimal
and high-signal — each item should direct AI attention to a distinct class of
real issues, not catalog every possible bug pattern.
Regular reviews use only Step 1 and never prompt. Steps 2–4 are an explicit
maintenance mode under `SKILL.md` § Interaction and interruption contract.
Interactive maintenance may request the declared selection or destination;
automated maintenance reports candidates or missing destination information,
applies no selection-dependent edits, and stops safely.
## State model
- **Proposed**: a valid candidate drafted from a recurring uncovered pattern
and recorded in the review report. No checklist file is edited. This state
is session-scoped: the report is not a durable carrier, so a candidate does
not survive into a later session.
- **Accepted**: a user explicitly selects a proposed candidate during a
separately authorized checklist-maintenance flow. Acceptance authorizes the
rule choice, not an implicit commit, push, or PR.
- **Persisted**: an accepted rule has been written to the canonical tracked
checklist in a user-designated long-lived checkout and captured in a durable
repository record such as a commit or PR. Only this state may be described as
available to subsequent reviews.
The canonical storage is the appropriate tracked checklist under
`.agents/skills/gh-pr-review/references/` — normally `code-checklist.md` or
`doc-checklist.md`. Never treat runtime memory, a report, or an uncommitted file
as canonical storage.
## Step 1: Draft proposed candidates
For each uncovered pattern, draft a candidate item. ALL rules below MUST be
satisfied — violation makes the candidate invalid:
1. One assertive phrase describing the expected state (not a question)
2. Generic: applies across files, not tied to a specific variable, function, or bug
3. Atomic: one checkable concern per item (not "X and Y")
4. No overlap: if the issue is a specific case of an existing item, do NOT add it
5. Place under the most specific existing category; create a new category only when
no existing one fits
6. Each category stays within 3–8 items. Below 3, merge into a related category.
Above 8, first try merging overlapping items; only split if each resulting
sub-category has a distinct focus expressible in 2–3 words
7. Prefer fewer, broader items — the checklist is a prompt for attention directions,
not an exhaustive bug catalog
When uncertain whether a new item overlaps with an existing one, do NOT add it.
During a regular local, teams, or PR review, include every valid candidate in
the final report as **Proposed**, with its target checklist and category. Stop
there: do not prompt for selection, edit a checklist, or claim persistence.
## Step 2: Accept in an explicit maintenance flow
Continue past Step 1 only when the user explicitly requests checklist
maintenance or asks to adopt proposed candidates.
**Proposed candidates live only in the session that drafted them.** A review
report is session output, not a durable store, so maintenance runs only in the
same session as the review that produced the candidates. If this session holds
no proposed candidates, say so and stop — do not reconstruct candidates from
memory, from an earlier session's report, or by re-reviewing. To adopt
candidates from an earlier session, re-run the review that produces them.
In an interactive run, present this session's candidates for selection.
Unselected candidates remain proposed or are discarded; selected candidates
become **Accepted**. In an automated run, report the candidates and stop before
acceptance.
Acceptance does not authorize writing into a temporary PR review worktree. The
maintenance flow must identify a user-designated persistent checkout and
target branch. If either is unclear, an interactive run asks before editing; an
automated run reports the missing destination and stops. Never guess.
## Step 3: Write accepted rules
Insert accepted items into the canonical checklist file in the designated
persistent checkout at the appropriate position under the category and
priority rules above. Never write them into `REVIEW_DIR` or another disposable
worktree. At this point the rules are accepted and written, but not yet
persisted.
## Step 4: Create the durable record
Commit, push, or open a PR only with the authority required by the repository's
normal publish workflow. Record the target branch and resulting commit or PR in
the maintenance report. After that durable record exists, mark the rules
**Persisted**. If publication was not authorized or fails, report them as
accepted but not persisted.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |