Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate or update project documentation verified against the codebase
.claude/skills/coco-research-gsd-docs-update/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -8% | 0% |
<objective> Generate and update up to 9 documentation files for the current project. Each doc type is written by a gsd-doc-writer subagent that explores the codebase directly — no hallucinated paths, phantom endpoints, or stale signatures.
Flag handling rule:
$ARGUMENTS$ARGUMENTS, treat it as inactive--force: skip preservation prompts, regenerate all docs regardless of existing content or GSD markers--verify-only: check existing docs for accuracy against codebase, no generation (full verification requires Phase 4 verifier)--force and --verify-only both appear in $ARGUMENTS, --force takes precedence</objective>
<execution_context> @$HOME/.claude/get-shit-done/workflows/docs-update.md </execution_context>
<context> Arguments: $ARGUMENTS
Available optional flags (documentation only — not automatically active):
--force — Regenerate all docs. Overwrites hand-written and GSD docs alike. No preservation prompts.--verify-only — Check existing docs for accuracy against the codebase. No files are written. Reports VERIFY marker count. Full codebase fact-checking requires the gsd-doc-verifier agent (Phase 4).Active flags must be derived from $ARGUMENTS:
--force is active only if the literal --force token is present in $ARGUMENTS--verify-only is active only if the literal --verify-only token is present in $ARGUMENTS</context>
<process> Execute the docs-update workflow from @$HOME/.claude/get-shit-done/workflows/docs-update.md end-to-end. Preserve all workflow gates (preservation_check, flag handling, wave execution, monorepo dispatch, commit, reporting). </process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→fail | 7,607 | 17,224 | +126% | 1 | 1 | 0% | 216 | 798 | +269% | 0 | 0 | — |
case-01 | fail→fail | 2,858 | 21,368 | +648% | 1 | 1 | 0% | 412 | 940 | +128% | 0 | 0 | — |
case-02 | fail→fail | 15,917 | 17,507 | +10% | 1 | 1 | 0% | 1,830 | 988 | -46% | 0 | 0 | — |
case-03 | fail→fail | 5,333 | 5,468 | +3% | 1 | 1 | 0% | 628 | 699 | +11% | 0 | 0 | — |
case-05 | fail→fail | 14,774 | 56,491 | +282% | 1 | 1 | 0% | 1,774 | 1,198 | -32% | 0 | 0 | — |
case-06 | pass→fail | 13,970 | 7,006 | -50% | 1 | 1 | 0% | 1,672 | 822 | -51% | 0 | 0 | — |
case-07 | fail→fail | 17,154 | 34,411 | +101% | 1 | 1 | 0% | 1,885 | 851 | -55% | 0 | 0 | — |
case-08 | pass→pass | 8,215 | 8,072 | -2% | 1 | 1 | 0% | 1,313 | 1,018 | -22% | 0 | 0 | — |
case-09 | fail→pass | 7,135 | 7,415 | +4% | 1 | 1 | 0% | 1,088 | 780 | -28% | 0 | 0 | — |
case-10 | fail→pass | 13,701 | 1,778 | -87% | 1 | 1 | 0% | 1,276 | 785 | -38% | 0 | 0 | — |
case-11 | fail→pass | 9,852 | 7,473 | -24% | 1 | 1 | 0% | 1,439 | 772 | -46% | 0 | 0 | — |
case-12 | fail→pass | 6,282 | 7,636 | +22% | 1 | 1 | 0% | 901 | 739 | -18% | 0 | 0 | — |
case-13 | fail→pass | 8,634 | 9,738 | +13% | 1 | 1 | 0% | 1,372 | 1,260 | -8% | 0 | 0 | — |
case-14 | fail→pass | 51,125 | 4,067 | -92% | 1 | 1 | 0% | 8,214 | 1,153 | -86% | 0 | 0 | — |
case-15 | pass→pass | 13,094 | 3,505 | -73% | 1 | 1 | 0% | 1,063 | 914 | -14% | 0 | 0 | — |
case-16 | fail→pass | 9,480 | 10,709 | +13% | 1 | 1 | 0% | 1,294 | 1,284 | -1% | 0 | 0 | — |
case-17 | fail→fail | 16,128 | 8,011 | -50% | 1 | 1 | 0% | 1,442 | 972 | -33% | 0 | 0 | — |
case-18 | fail→fail | 14,782 | 6,742 | -54% | 1 | 1 | 0% | 2,289 | 1,487 | -35% | 0 | 0 | — |
case-19 | pass→pass | 14,435 | 7,965 | -45% | 1 | 1 | 0% | 2,055 | 1,030 | -50% | 0 | 0 | — |
case-20 | pass→fail | 19,398 | 8,843 | -54% | 1 | 1 | 0% | 2,416 | 963 | -60% | 0 | 0 | — |
case-21 | pass→pass | 11,630 | 7,958 | -32% | 1 | 1 | 0% | 1,067 | 935 | -12% | 0 | 0 | — |
case-22 | pass→pass | 7,926 | 3,618 | -54% | 1 | 1 | 0% | 357 | 905 | +154% | 0 | 0 | — |
case-23 | fail→fail | 12,892 | 9,677 | -25% | 1 | 1 | 0% | 1,322 | 915 | -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. 23 cases were attempted, and 14 counted toward the lift figure. The other 9 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 +22 percentage points is the difference between those two pass rates over the 14 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.