Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Discover tech stack.
.claude/skills/griddynamics-init-workspace-discovery/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -18% | 0% |
<init_workspace_discovery>
<role> Senior workspace cartographer — fast, factual technical inventory. </role>
<when_to_use_skill>
Without factual inventory of tech stack, structure, and dependencies, subsequent phases operate blind. Use during workspace initialization or when TECHSTACK, CODEMAP, or DEPENDENCIES are missing or stale.
</when_to_use_skill>
<process>
agents/TEMP/ foldergit ls-files --cached --others --exclude-standard in each repository or fallback to find/ls/etc with filtersMinimal set must be present: ... # Rosetta agents/TEMP/ refsrc/ !refsrc/INDEX.md
</process>
<files>
</files>
<pitfalls>
</pitfalls>
<references>
Example scripts provided (think if you want to use it, as those are very large, 20K each, use ACQUIRE FROM KB command to load):
init-workspace-discovery/scripts/codemap.ps1.txt init-workspace-discovery/scripts/codemap.sh.txt NOTE: .txt extension is added to avoid execution or treating as executable.
</references>
</init_workspace_discovery>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,319 | 4,454 | +34% | 1 | 1 | 0% | 197 | 947 | +381% | 0 | 0 | — |
case-02 | fail→fail | 4,424 | 4,954 | +12% | 1 | 1 | 0% | 275 | 925 | +236% | 0 | 0 | — |
case-03 | fail→fail | 2,446 | 4,475 | +83% | 1 | 1 | 0% | 350 | 910 | +160% | 0 | 0 | — |
case-04 | fail→pass | 8,916 | 1,360 | -85% | 1 | 1 | 0% | 1,454 | 893 | -39% | 0 | 0 | — |
case-05 | fail→pass | 10,424 | 1,827 | -82% | 1 | 1 | 0% | 1,616 | 985 | -39% | 0 | 0 | — |
case-06 | pass→pass | 3,951 | 3,888 | -2% | 1 | 1 | 0% | 692 | 1,301 | +88% | 0 | 0 | — |
case-07 | fail→pass | 4,981 | 2,271 | -54% | 1 | 1 | 0% | 775 | 1,049 | +35% | 0 | 0 | — |
case-08 | pass→pass | 6,101 | 3,144 | -48% | 1 | 1 | 0% | 942 | 1,099 | +17% | 0 | 0 | — |
case-09 | pass→pass | 11,245 | 3,383 | -70% | 1 | 1 | 0% | 1,631 | 1,231 | -25% | 0 | 0 | — |
case-10 | pass→pass | 6,081 | 1,902 | -69% | 1 | 1 | 0% | 1,033 | 1,016 | -2% | 0 | 0 | — |
case-11 | fail→pass | 10,703 | 3,264 | -70% | 1 | 1 | 0% | 1,697 | 1,178 | -31% | 0 | 0 | — |
case-12 | fail→pass | 9,984 | 3,750 | -62% | 1 | 1 | 0% | 1,691 | 1,390 | -18% | 0 | 0 | — |
case-13 | pass→pass | 15,141 | 2,445 | -84% | 1 | 1 | 0% | 2,419 | 1,075 | -56% | 0 | 0 | — |
case-14 | fail→pass | 9,533 | 3,254 | -66% | 1 | 1 | 0% | 1,587 | 1,204 | -24% | 0 | 0 | — |
case-15 | pass→pass | 7,527 | 2,447 | -67% | 1 | 1 | 0% | 1,066 | 1,126 | +6% | 0 | 0 | — |
case-16 | pass→pass | 13,174 | 2,359 | -82% | 1 | 1 | 0% | 2,305 | 1,101 | -52% | 0 | 0 | — |
case-17 | pass→pass | 12,735 | 3,831 | -70% | 1 | 1 | 0% | 1,933 | 1,259 | -35% | 0 | 0 | — |
case-18 | pass→pass | 14,959 | 5,852 | -61% | 1 | 1 | 0% | 2,215 | 1,673 | -24% | 0 | 0 | — |
case-19 | pass→pass | 9,399 | 3,369 | -64% | 1 | 1 | 0% | 1,529 | 1,197 | -22% | 0 | 0 | — |
case-20 | fail→pass | 10,049 | 4,198 | -58% | 1 | 1 | 0% | 1,514 | 1,308 | -14% | 0 | 0 | — |
case-21 | fail→fail | 10,557 | 4,370 | -59% | 1 | 1 | 0% | 1,774 | 1,361 | -23% | 0 | 0 | — |
case-22 | pass→pass | 10,601 | 2,820 | -73% | 1 | 1 | 0% | 1,645 | 1,108 | -33% | 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. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 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 19 comparable cases.
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.