From team-code-review
The team's code review rubric — what to check for correctness, security, style/conventions, and performance, plus the expected output format. Use when reviewing code changes, a diff, or a set of files.
How this skill is triggered — by the user, by Claude, or both
Slash command
/team-code-review:review-rubricThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This is the team's shared, authoritative review standard. Edit it via pull
This is the team's shared, authoritative review standard. Edit it via pull
request — every change to how we review lives here, so both the automatic
hook and the manual /review command stay in sync.
Review the change, not the whole codebase. Prefer reviewing a diff. Verify claims against the code; do not assume the author got it right.
catch {})? Are errors actionable?await.eval/exec on untrusted data.Start with a one-line verdict: APPROVE / APPROVE WITH NITS / CHANGES REQUESTED.
Then findings grouped by severity, each line:
severity —
path:line— problem, why it matters, and a concrete fix.
Severity:
Keep it tight and specific. Cite file:line. A short accurate review beats a
long padded one — don't manufacture findings.
Add team-specific rules under the relevant section above — e.g. "all DB access
goes through repo/", "no console.log in packages/server", "prefer
Result<T> over throwing in core/". Keep entries concrete and checkable so
the reviewer can actually verify them.
npx claudepluginhub crazywillbear/my-dotclaude --plugin workflowProvides CDSS development patterns for drug interaction checking, dose validation, clinical scoring (NEWS2, qSOFA), and alert classification integrated into EMR workflows.