Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Walks every markdown file in the knowledge base, extracts metadata and content, and rebuilds the SQLite FTS5 search index. Run after major file changes (bulk edits, reorganizations, imports).
.claude/skills/miosa-osa-index/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 186% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -82% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -79% | 0% |
> Rebuild the full-text search index across all knowledge base files.
/index [--full] [--stats]Walks every markdown file in the knowledge base, extracts metadata and content, and rebuilds the SQLite FTS5 search index. Run after major file changes (bulk edits, reorganizations, imports).
Runs: cd engine && mix optimal.index
Process:
With --stats: cd engine && mix optimal.stats to show index statistics.
bash# Full reindex /index # Reindex and show statistics /index --stats
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 3,196 | 12,686 | +297% | 1 | 1 | 0% | 472 | 1,349 | +186% | 0 | 0 | — |
case-03 | fail→pass | 7,204 | 7,655 | +6% | 1 | 1 | 0% | 1,057 | 868 | -18% | 0 | 0 | — |
case-04 | fail→pass | 15,458 | 2,295 | -85% | 1 | 1 | 0% | 2,477 | 450 | -82% | 0 | 0 | — |
case-01 | fail→pass | 7,718 | 4,016 | -48% | 1 | 1 | 0% | 1,003 | 790 | -21% | 0 | 0 | — |
case-05 | fail→pass | 15,062 | 2,413 | -84% | 1 | 1 | 0% | 2,485 | 532 | -79% | 0 | 0 | — |
case-06 | fail→pass | 8,877 | 3,485 | -61% | 1 | 1 | 0% | 1,235 | 615 | -50% | 0 | 0 | — |
case-07 | pass→pass | 8,927 | 2,095 | -77% | 1 | 1 | 0% | 1,360 | 495 | -64% | 0 | 0 | — |
case-08 | fail→pass | 8,000 | 2,168 | -73% | 1 | 1 | 0% | 1,251 | 526 | -58% | 0 | 0 | — |
case-14 | fail→pass | 11,787 | 2,494 | -79% | 1 | 1 | 0% | 1,627 | 481 | -70% | 0 | 0 | — |
case-09 | fail→pass | 9,529 | 2,280 | -76% | 1 | 1 | 0% | 1,392 | 545 | -61% | 0 | 0 | — |
case-10 | pass→pass | 11,933 | 2,702 | -77% | 1 | 1 | 0% | 1,834 | 664 | -64% | 0 | 0 | — |
case-11 | pass→pass | 11,555 | 3,287 | -72% | 1 | 1 | 0% | 1,758 | 642 | -63% | 0 | 0 | — |
case-12 | fail→pass | 11,254 | 2,620 | -77% | 1 | 1 | 0% | 1,913 | 588 | -69% | 0 | 0 | — |
case-13 | fail→pass | 9,573 | 2,363 | -75% | 1 | 1 | 0% | 1,473 | 516 | -65% | 0 | 0 | — |
case-15 | fail→pass | 12,930 | 1,853 | -86% | 1 | 1 | 0% | 2,121 | 468 | -78% | 0 | 0 | — |
case-16 | fail→pass | 12,359 | 2,575 | -79% | 1 | 1 | 0% | 1,807 | 570 | -68% | 0 | 0 | — |
case-17 | pass→pass | 8,494 | 2,492 | -71% | 1 | 1 | 0% | 1,299 | 607 | -53% | 0 | 0 | — |
case-18 | pass→pass | 11,065 | 3,558 | -68% | 1 | 1 | 0% | 1,985 | 749 | -62% | 0 | 0 | — |
case-19 | fail→pass | 11,699 | 3,382 | -71% | 1 | 1 | 0% | 1,375 | 769 | -44% | 0 | 0 | — |
case-20 | fail→pass | 9,468 | 5,436 | -43% | 1 | 1 | 0% | 1,421 | 966 | -32% | 0 | 0 | — |
case-21 | fail→pass | 9,257 | 7,031 | -24% | 1 | 1 | 0% | 1,491 | 1,127 | -24% | 0 | 0 | — |
case-22 | pass→pass | 6,670 | 4,071 | -39% | 1 | 1 | 0% | 1,192 | 815 | -32% | 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 +73 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.