Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when locating existing project features, flows, pages, modules, components, utilities, patterns, contracts, business concepts, or similar functionality before answering where something is or before building related functionality. Trigger for 项目里有没有做过, 在哪里, 类似功能, 已有功能, 复用, 不要重复造轮子, 项目里叫什么, 哪些文件, 找一下, existing feature, similar functionality, where is, reuse, avoid rebuilding, project concept, feature inventory, capability map, or implementation already exists. Do not use for pure styling, isol
.claude/skills/hashgraph-online-project-map-find/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | 101% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 31% | 0% |
Use .wingman/project-map/ to narrow candidate capabilities, then verify source before strong claims. The project map is a discovery map, not final authority.
Read indexes first, select a small candidate set, and verify source before strong claims. Do not turn a project-map lookup into a full repository scan unless the map is missing or clearly insufficient.
Continue only when the user needs to locate an existing capability, compare similar implementations, map a project concept, or check for reuse/reference candidates. If the task is only editing, explaining, testing, or styling and does not ask about existing capability discovery, do not use this skill.
Answer questions like:
Locate: find where something is.Compare: compare similar capabilities.Reuse Decision: decide reuse/extend/wrap/reference/create.Concept Mapping: map user language to project code names.Implementation Prep: find precedents before building..wingman/project-map/ is missing, say no project map exists yet.project-map-catalog..wingman/project-map/index.md when present..wingman/project-map/glossary/index.md when present.features, flows, surfaces, components, modules, utilities, patterns, contracts, or domains.relationships/index.md only when it exists and the query needs cross-entry relationships.references/candidate-ranking.md.Use these decision labels:
Locate: found where it lives, not necessarily reusable.Reuse: directly use existing implementation.Extend: expand the same semantic responsibility.Wrap: wrap existing capability with context, defaults, permissions, or layout.Reference: follow the flow/pattern but do not reuse code directly.Avoid: similar but should not be used.Create New: no suitable candidate.Catalog Needed: source search found a capability not in the project map.Put the strongest answer first:
markdownFound: Yes | Partial | No Closest Matches: - `[Entry Name]`: [what it does] - Where: `[path/route]` - Traits: [key traits] - Decision: Locate | Reuse | Extend | Wrap | Reference | Avoid | Create New | Catalog Needed - Evidence: [entry/source] Recommended Next Step: [read source / reuse module / create new / catalog missing feature] Gaps: [missing entry, stale verification, no map coverage]
If no good match exists, say so directly. Distinguish "project map has no match" from "source search found no match".
references/candidate-ranking.md only when the index yields multiple plausible candidates or the best candidate is unclear.references/output-examples.md only when shaping a complex final response.Before finishing, verify:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 2,375 | 15,586 | +556% | 1 | 1 | 0% | 306 | 1,532 | +401% | 0 | 0 | — |
case-02 | fail→fail | 27,817 | 5,757 | -79% | 1 | 1 | 0% | 3,936 | 1,530 | -61% | 0 | 0 | — |
case-03 | fail→fail | 15,316 | 16,193 | +6% | 1 | 1 | 0% | 251 | 1,557 | +520% | 0 | 0 | — |
case-04 | pass→pass | 14,912 | 12,534 | -16% | 1 | 1 | 0% | 1,365 | 2,945 | +116% | 0 | 0 | — |
case-05 | pass→fail | 26,621 | 16,647 | -37% | 1 | 1 | 0% | 2,964 | 1,598 | -46% | 0 | 0 | — |
case-06 | pass→fail | 14,968 | 17,206 | +15% | 1 | 1 | 0% | 1,944 | 1,547 | -20% | 0 | 0 | — |
case-07 | fail→fail | 15,496 | 9,366 | -40% | 1 | 1 | 0% | 271 | 1,494 | +451% | 0 | 0 | — |
case-08 | fail→fail | 14,466 | 11,612 | -20% | 1 | 1 | 0% | 2,481 | 1,515 | -39% | 0 | 0 | — |
case-09 | fail→fail | 24,882 | 14,991 | -40% | 1 | 1 | 0% | 2,738 | 1,368 | -50% | 0 | 0 | — |
case-10 | fail→fail | 14,493 | 14,881 | +3% | 1 | 1 | 0% | 1,950 | 1,395 | -28% | 0 | 0 | — |
case-11 | fail→pass | 15,108 | 12,966 | -14% | 1 | 1 | 0% | 1,342 | 2,696 | +101% | 0 | 0 | — |
case-12 | fail→pass | 14,687 | 8,902 | -39% | 1 | 1 | 0% | 1,434 | 1,618 | +13% | 0 | 0 | — |
case-13 | fail→pass | 10,474 | 15,026 | +43% | 1 | 1 | 0% | 1,741 | 2,652 | +52% | 0 | 0 | — |
case-14 | pass→pass | 14,807 | 9,490 | -36% | 1 | 1 | 0% | 1,532 | 1,784 | +16% | 0 | 0 | — |
case-15 | fail→pass | 19,224 | 15,253 | -21% | 1 | 1 | 0% | 1,960 | 2,315 | +18% | 0 | 0 | — |
case-16 | fail→pass | 16,223 | 10,482 | -35% | 1 | 1 | 0% | 1,481 | 1,947 | +31% | 0 | 0 | — |
case-17 | pass→pass | 12,229 | 7,452 | -39% | 1 | 1 | 0% | 1,231 | 1,562 | +27% | 0 | 0 | — |
case-18 | fail→pass | 14,727 | 11,338 | -23% | 1 | 1 | 0% | 2,352 | 2,043 | -13% | 0 | 0 | — |
case-19 | pass→pass | 6,691 | 7,538 | +13% | 1 | 1 | 0% | 857 | 1,486 | +73% | 0 | 0 | — |
case-20 | pass→pass | 8,470 | 7,382 | -13% | 1 | 1 | 0% | 1,433 | 1,483 | +3% | 0 | 0 | — |
case-21 | fail→fail | 13,897 | 5,865 | -58% | 1 | 1 | 0% | 1,402 | 1,373 | -2% | 0 | 0 | — |
case-22 | fail→pass | 39,526 | 12,210 | -69% | 1 | 1 | 0% | 2,272 | 2,503 | +10% | 0 | 0 | — |
case-23 | pass→pass | 10,668 | 1,948 | -82% | 1 | 1 | 0% | 1,325 | 1,501 | +13% | 0 | 0 | — |
case-24 | pass→pass | 18,818 | 7,855 | -58% | 1 | 1 | 0% | 2,128 | 2,622 | +23% | 0 | 0 | — |
case-25 | fail→pass | 13,977 | 1,289 | -91% | 1 | 1 | 0% | 1,841 | 1,421 | -23% | 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 15 counted toward the lift figure. The other 10 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 +24 percentage points is the difference between those two pass rates over the 15 comparable cases. 3 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.