Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →notque/parallel-code-review
Parallel 3-reviewer code review: Security, Business-Logic, Architecture.
Passed 15 of 19 test cases with the skill, and 9 of 19 without it.
and 5 more in the full report
8 fixed9 passed either way4 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
This benchmark measured an earlier version of the skill.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
During a review aggregation phase, the team has accumulated 4 total findings across all reviewers: 1 MEDIUM finding from Security, 2 MEDIUM findings from Business Logic, and 1 LOW finding from Architecture. Should the adversarial verification phase be executed or skipped?
| Checked | Without skill | With skill |
|---|---|---|
| Executes the adversarial verification pass when total findings reach four. | Missed | Passed |
The instructions your model reads when the skill loads.
Divide a review among Security, Business Logic, and Architecture reviewers, then combine their findings. Use the shared procedure in ../systematic-code-review/SKILL.md for evidence, finding verification, and review reuse.
| Task | Reference |
|---|---|
| Select PR review scope and roster | ../../process/pr-workflow/references/pr-risk-policy.md |
| Brief the Architecture reviewer | references/architecture-smell-baseline.md |
Read repository instructions and identify the requested diff or files:
git diff --name-only HEAD~1
gh pr view --json files -q '.files[].path'
Use the actual requested base, not HEAD~1, for a multi-commit change. Record the reviewed revision and relevant existing checks.
Honor an explicit user or /do roster. For this skill's three-reviewer mode, keep all three roles. Do not start another trio when this skill is used inside an already selected roster. Automatic PR review selection belongs to the risk policy, including its high-risk requirements.
Select the Architecture reviewer by language: Go → golang-general-engineer; Python → python-general-engineer; TypeScript → typescript-frontend-engineer; mixed or other → Explore. Add threat modeling, git bisect, or benchmarks when explicitly requested.
Gate: Scope, roster, and review evidence are known.
Dispatch the selected independent reviewers together. Reviewers read and report; they do not edit code. Supply the same task intent, diff base/head, constraints, and relevant check results, with a distinct focus:
| Role | Focus |
|---|---|
| Security | Authentication, authorization, input validation, secrets, relevant OWASP risks |
| Business Logic | Requirements, edge cases, state transitions, validation, failure modes, documentation accuracy |
| Architecture | Design, structure, performance, maintainability, scope, simplicity |
All reviewers can flag documentation errors and scope creep. Give the Architecture reviewer references/architecture-smell-baseline.md verbatim; it contains the 12 smells, language counter-examples, and severity cap.
test cases passed, vs 9 / 19 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.