Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Review source files changed during a phase for bugs, security issues, and code quality problems
.claude/skills/coco-research-gsd-code-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -26% | 0% |
<objective> Review source files changed during a phase for bugs, security vulnerabilities, and code quality problems.
Spawns the gsd-code-reviewer agent to analyze code at the specified depth level. Produces REVIEW.md artifact in the phase directory with severity-classified findings.
Arguments:
--depth=quick|standard|deep (optional) — review depth level, overrides workflow.code_review_depth config--files file1,file2,... (optional) — explicit comma-separated file list, skips SUMMARY/git scoping (highest precedence for scoping)Output: {padded_phase}-REVIEW.md in phase directory + inline summary of findings </objective>
<execution_context> @$HOME/.claude/get-shit-done/workflows/code-review.md </execution_context>
<context> Phase: $ARGUMENTS (first positional argument is phase number)
Optional flags parsed from $ARGUMENTS:
--depth=VALUE — Depth override (quick|standard|deep). If provided, overrides workflow.code_review_depth config.--files=file1,file2,... — Explicit file list override. Has highest precedence for file scoping. When provided, workflow skips SUMMARY.md extraction and git diff fallback entirely.Context files (CLAUDE.md, SUMMARY.md, phase state) are resolved inside the workflow via gsd-tools init phase-op and delegated to agent via <files_to_read> blocks. </context>
<process> This command is a thin dispatch layer. It parses arguments and delegates to the workflow.
Execute the code-review workflow from @$HOME/.claude/get-shit-done/workflows/code-review.md end-to-end.
The workflow (not this command) enforces these gates:
</process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 20,281 | 18,686 | -8% | 1 | 1 | 0% | 2,249 | 2,092 | -7% | 0 | 0 | — |
case-17 | fail→pass | 16,740 | 14,368 | -14% | 1 | 1 | 0% | 1,521 | 1,730 | +14% | 0 | 0 | — |
case-02 | fail→fail | 13,324 | 11,696 | -12% | 1 | 1 | 0% | 857 | 969 | +13% | 0 | 0 | — |
case-03 | fail→fail | 8,271 | 16,442 | +99% | 1 | 1 | 0% | 1,407 | 930 | -34% | 0 | 0 | — |
case-04 | pass→fail | 9,723 | 17,893 | +84% | 1 | 1 | 0% | 622 | 1,101 | +77% | 0 | 0 | — |
case-05 | fail→fail | 7,649 | 16,628 | +117% | 1 | 1 | 0% | 377 | 947 | +151% | 0 | 0 | — |
case-06 | fail→fail | 11,687 | 23,292 | +99% | 1 | 1 | 0% | 1,114 | 912 | -18% | 0 | 0 | — |
case-07 | fail→fail | 16,184 | 8,021 | -50% | 1 | 1 | 0% | 1,759 | 1,022 | -42% | 0 | 0 | — |
case-08 | fail→fail | 12,881 | 17,846 | +39% | 1 | 1 | 0% | 1,952 | 1,042 | -47% | 0 | 0 | — |
case-09 | fail→fail | 11,342 | 12,985 | +14% | 1 | 1 | 0% | 1,989 | 895 | -55% | 0 | 0 | — |
case-10 | fail→pass | 8,126 | 9,717 | +20% | 1 | 1 | 0% | 1,432 | 958 | -33% | 0 | 0 | — |
case-11 | pass→pass | 13,556 | 7,838 | -42% | 1 | 1 | 0% | 1,232 | 1,043 | -15% | 0 | 0 | — |
case-12 | pass→pass | 3,720 | 2,243 | -40% | 1 | 1 | 0% | 678 | 905 | +33% | 0 | 0 | — |
case-13 | fail→pass | 20,287 | 2,695 | -87% | 1 | 1 | 0% | 2,610 | 957 | -63% | 0 | 0 | — |
case-14 | fail→fail | 13,043 | 6,874 | -47% | 1 | 1 | 0% | 1,041 | 711 | -32% | 0 | 0 | — |
case-15 | pass→pass | 10,209 | 6,274 | -39% | 1 | 1 | 0% | 1,568 | 1,277 | -19% | 0 | 0 | — |
case-16 | fail→pass | 14,015 | 3,453 | -75% | 1 | 1 | 0% | 1,179 | 1,028 | -13% | 0 | 0 | — |
case-18 | fail→pass | 9,110 | 11,514 | +26% | 1 | 1 | 0% | 1,424 | 1,053 | -26% | 0 | 0 | — |
case-19 | pass→pass | 20,317 | 8,426 | -59% | 1 | 1 | 0% | 2,502 | 1,089 | -56% | 0 | 0 | — |
case-20 | fail→pass | 8,411 | 5,069 | -40% | 1 | 1 | 0% | 995 | 1,103 | +11% | 0 | 0 | — |
case-21 | fail→pass | 9,427 | 6,768 | -28% | 1 | 1 | 0% | 1,579 | 793 | -50% | 0 | 0 | — |
case-22 | pass→pass | 10,778 | 7,370 | -32% | 1 | 1 | 0% | 899 | 961 | +7% | 0 | 0 | — |
case-23 | pass→pass | 16,052 | 6,931 | -57% | 1 | 1 | 0% | 1,858 | 800 | -57% | 0 | 0 | — |
case-24 | fail→pass | 17,114 | 6,962 | -59% | 1 | 1 | 0% | 1,963 | 848 | -57% | 0 | 0 | — |
case-25 | fail→pass | 14,125 | 11,219 | -21% | 1 | 1 | 0% | 2,363 | 1,637 | -31% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted, and 18 counted toward the lift figure. The other 7 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +32 percentage points is the difference between those two pass rates over the 18 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.