Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Look up Godot engine class APIs — methods, properties, signals, enums. Use when you need to find which class to use or look up specific API details. Supports version-specific docs — reads the correct API for the project's Godot version.
.claude/skills/randallliuxin-godot-api/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -77% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -74% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -72% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -52% | 0% |
$ARGUMENTS
Docs are stored per Godot version under doc_api/{version}/ and doc_api_csharp/{version}/.
Determine the version to use:
project.godot exists in the working directory, extract from config/featureslatestBootstrap if docs for the target version are missing:
bashbash ${CLAUDE_SKILL_DIR}/tools/ensure_doc_api.sh [version]
Language selection: doc_api/{version}/ contains GDScript docs. doc_api_csharp/{version}/ contains C# docs. Default to GDScript unless the caller asks about C#.
${CLAUDE_SKILL_DIR}/doc_api/{version}/_common.md — index of ~128 common classes${CLAUDE_SKILL_DIR}/doc_api_csharp/{version}/_common.md_other.md in the same directory${CLAUDE_SKILL_DIR}/doc_api/{version}/{ClassName}.md for full API${CLAUDE_SKILL_DIR}/doc_api_csharp/{version}/{ClassName}.mdC# syntax reference: ${CLAUDE_SKILL_DIR}/csharp.md — C# Godot syntax, patterns, and recipes.
GDScript syntax reference: ${CLAUDE_SKILL_DIR}/gdscript.md — GDScript language spec, type system, patterns.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 4,857 | 2,655 | -45% | 1 | 1 | 0% | 661 | 840 | +27% | 0 | 0 | — |
case-02 | pass→pass | 15,028 | 1,483 | -90% | 1 | 1 | 0% | 2,500 | 637 | -75% | 0 | 0 | — |
case-03 | pass→pass | 7,967 | 1,608 | -80% | 1 | 1 | 0% | 1,326 | 605 | -54% | 0 | 0 | — |
case-04 | fail→pass | 19,348 | 1,977 | -90% | 1 | 1 | 0% | 2,973 | 689 | -77% | 0 | 0 | — |
case-05 | fail→pass | 19,086 | 2,288 | -88% | 1 | 1 | 0% | 3,024 | 797 | -74% | 0 | 0 | — |
case-06 | fail→pass | 7,479 | 1,689 | -77% | 1 | 1 | 0% | 1,010 | 645 | -36% | 0 | 0 | — |
case-07 | fail→pass | 16,265 | 2,052 | -87% | 1 | 1 | 0% | 2,603 | 741 | -72% | 0 | 0 | — |
case-08 | fail→pass | 9,020 | 2,132 | -76% | 1 | 1 | 0% | 1,537 | 743 | -52% | 0 | 0 | — |
case-13 | fail→pass | 11,000 | 2,404 | -78% | 1 | 1 | 0% | 1,795 | 813 | -55% | 0 | 0 | — |
case-09 | fail→pass | 10,992 | 1,417 | -87% | 1 | 1 | 0% | 1,620 | 586 | -64% | 0 | 0 | — |
case-10 | fail→pass | 6,616 | 1,682 | -75% | 1 | 1 | 0% | 1,031 | 692 | -33% | 0 | 0 | — |
case-11 | fail→pass | 5,790 | 1,676 | -71% | 1 | 1 | 0% | 866 | 629 | -27% | 0 | 0 | — |
case-12 | fail→pass | 5,884 | 2,126 | -64% | 1 | 1 | 0% | 924 | 722 | -22% | 0 | 0 | — |
case-14 | pass→pass | 9,031 | 2,966 | -67% | 1 | 1 | 0% | 1,429 | 869 | -39% | 0 | 0 | — |
case-15 | pass→pass | 14,500 | 6,291 | -57% | 1 | 1 | 0% | 2,482 | 1,421 | -43% | 0 | 0 | — |
case-16 | pass→pass | 5,030 | 1,968 | -61% | 1 | 1 | 0% | 826 | 777 | -6% | 0 | 0 | — |
case-17 | fail→pass | 12,295 | 2,700 | -78% | 1 | 1 | 0% | 2,232 | 921 | -59% | 0 | 0 | — |
case-18 | fail→pass | 5,473 | 1,964 | -64% | 1 | 1 | 0% | 873 | 812 | -7% | 0 | 0 | — |
case-19 | fail→pass | 7,266 | 1,992 | -73% | 1 | 1 | 0% | 1,166 | 752 | -36% | 0 | 0 | — |
case-20 | pass→pass | 13,252 | 9,998 | -25% | 1 | 1 | 0% | 2,182 | 1,990 | -9% | 0 | 0 | — |
case-21 | pass→pass | 12,577 | 8,609 | -32% | 1 | 1 | 0% | 2,215 | 1,956 | -12% | 0 | 0 | — |
case-22 | pass→pass | 16,149 | 10,591 | -34% | 1 | 1 | 0% | 2,772 | 2,340 | -16% | 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. The headline lift of +59 percentage points is the difference between those two pass rates over the 22 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.