* chore: remove wm-cursor, add local-review skill, update PR skill for EE - Remove the unused wm-cursor script and all references to it in README_WORKMUX_DEV.md and worktree-common.sh - Add /local-review skill for code review (bugs + CLAUDE.md compliance) - Add EE companion PR workflow to the /pr skill Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: add wm-ts-nav tree-sitter navigator and fix format hooks - Add wm-ts-nav: standalone tree-sitter code navigator with SQLite index for fast symbol search, definition lookup, and file outlines across Rust, TypeScript, and Svelte files (~12ms warm, ~1s cold for 482 files) - Fix format hooks: surface errors instead of swallowing with 2>/dev/null, use direct prettier path with svelte plugin, add success feedback - Add wm-ts-nav commands to settings allow list - Document wm-ts-nav usage in CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * feat(wm-ts-nav): add refs command and --parent filter - refs: find usages of a symbol in code, skipping comments and strings (tree-sitter AST walk, ~46ms for 482 files vs grep's 4ms but no noise) - --parent filter on search: find all methods on a type across all files (e.g. search "%" --kind function --parent ServiceName) - Update CLAUDE.md with clearer when-to-use guidance Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * feat(wm-ts-nav): index refs in DB with import-path resolution Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * feat(wm-ts-nav): add body, callers, callees commands and refs --file/--caller - body: extract a symbol's source code from disk using indexed line ranges - callers: cross-file call graph via SQL join of refs + symbols tables - callees: list all identifiers referenced within a symbol's body - refs --file: scope results to files matching a substring - refs --caller: annotate each ref with the containing function name Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * feat(wm-ts-nav): add auto-rebuilding wrapper script The `wm-ts-nav/nav` wrapper checks if source files are newer than the binary and rebuilds automatically. Invoked via `sh wm-ts-nav/nav` to avoid needing executable permissions after clone. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: tighten CLAUDE.md nav section for actionable guidance Remove redundant question→command mapping, latency numbers, and excessive examples. Lead with "prefer wm-ts-nav over Read to save context window" and keep only the patterns that change behavior. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: revert backend/Cargo.lock to main Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: promote wm-ts-nav in workflow, copy binary to worktrees - CLAUDE.md: integrate wm-ts-nav into Workflow step 1 and Core Principles so agents use outline/body before full file reads - workmux: copy built binary via files.copy - worktree-common.sh: copy binary in wm_copy_dependencies for webmux Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(wm-ts-nav): fix double indexing, add TSX grammar, remove needless clone - Reuse index stats from the pre-query update instead of indexing twice on the Index command - Add Lang::Tsx variant so .tsx/.jsx files use LANGUAGE_TSX instead of LANGUAGE_TYPESCRIPT (Svelte stays on TS since script blocks are pure TS) - Remove source.clone() for non-Svelte files — move directly instead Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(wm-ts-nav): fix svelte line numbers, add class methods, innermost caller Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
3.2 KiB
3.2 KiB
name, user_invocable, description
| name | user_invocable | description |
|---|---|---|
| local-review | true | Code review a pull request for bugs and CLAUDE.md compliance. MUST use when asked to review code. |
Local Code Review Skill
Review a pull request for real bugs and CLAUDE.md compliance violations. This review targets HIGH SIGNAL issues only.
Review Philosophy
- Only flag issues you are certain about. If you are not sure an issue is real, do not flag it. False positives erode trust and waste reviewer time.
- Think like a senior engineer doing a final review — flag things that would cause incidents, not things that are merely imperfect.
What to Flag
- Code that won't compile or parse (syntax errors, type errors, missing imports)
- Code that will definitely produce wrong results regardless of inputs
- Clear, unambiguous CLAUDE.md violations (quote the exact rule being violated)
- Security issues in introduced code (injection, auth bypass, data exposure)
- Incorrect logic that will fail in production
What NOT to Flag
- Code style or quality concerns
- Potential issues that depend on specific inputs or runtime state
- Subjective suggestions or improvements
- Pre-existing issues not introduced by this PR
- Pedantic nitpicks a senior engineer wouldn't flag
- Issues a linter or type checker will catch
- General quality concerns unless explicitly prohibited in CLAUDE.md
- Issues silenced via lint ignore comments
Execution Steps
-
Determine the PR scope:
- If an argument is provided, use it as the PR number or branch
- Otherwise, detect from the current branch vs main
- Run
gh pr viewif a PR exists, or usegit diff main...HEAD
-
Find relevant CLAUDE.md files:
- Read the root
CLAUDE.md - Check for CLAUDE.md files in directories containing changed files
- Read the root
-
Get the diff and metadata:
gh pr difforgit diff main...HEADfor the full diffgh pr vieworgit log main..HEAD --onelinefor context
-
Read changed files where the diff alone is insufficient to understand context
-
Review for:
- CLAUDE.md compliance — check each rule against the changed code
- Bugs and logic errors — will this code work correctly?
- Security issues — injection, auth, data exposure in new code
-
Self-validate each finding: Before reporting, ask yourself:
- "Is this definitely a real issue, not a false positive?"
- "Would a senior engineer flag this in review?"
- If the answer to either is no, discard the finding
-
Output findings to the terminal (default) or post as PR comments (with
--commentflag)
Output Format
## Code review
Found N issues:
1. <description> (<reason: CLAUDE.md adherence | bug | security>)
<file_path:line_number>
2. <description> (<reason>)
<file_path:line_number>
If no issues are found:
## Code review
No issues found. Checked for bugs and CLAUDE.md compliance.
Posting Comments (--comment flag)
If the user passes --comment, post findings as inline PR comments using:
gh pr review --comment --body "<summary>"
Or for inline comments on specific lines:
gh api repos/{owner}/{repo}/pulls/{pr}/reviews -f body="<summary>" -f event="COMMENT" -f comments="[...]"