From engineering
Re-read the code you just wrote (and anything else you modified) with fresh eyes and look hard for bugs, errors, awkward design, dead code, or things that will confuse the next reader. Then fix what you find. Use after a batch of edits, when the user says "careful review", "look it over again", "give it another pass", or before opening a PR.
How this skill is triggered — by the user, by Claude, or both
Slash command
/engineering:careful-reviewThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
A second-look review of code you (or another agent) just wrote, with the goal of finding the bugs and confusions a first pass missed.
A second-look review of code you (or another agent) just wrote, with the goal of finding the bugs and confusions a first pass missed.
Phase 1 — Re-read everything you touched.
Walk every file you changed in this session, in order, and read the full diff plus the surrounding context. Don't just skim the changed lines: read the function they live in, read the callers, read the tests. The bugs are usually at the seams between what you changed and what you didn't.
For each change, ask:
Phase 2 — Surface findings before fixing.
Don't silently fix everything. Group the findings by file and rank by severity:
Present the list, pick a path with the user if anything is ambiguous, then fix.
Phase 3 — Fix and verify.
For each finding you're addressing, make the change, then re-read the diff one more time. Run the build, type-check, and tests after each batch of fixes — not at the end. A careful-review pass that introduces new bugs is worse than no pass.
Skip if the session change was trivial (a one-line rename, a comment fix). Run it after anything substantive: new features, refactors, bug fixes, anything across more than two files.
/engineering:code-review — pre-merge review of a PR you didn't author./engineering:codex-review — second opinion from OpenAI Codex against a base ref./engineering:security-review — security-focused pass on the same code.npx claudepluginhub goldenberry-so/pace --plugin engineeringGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.