Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build or update the code review knowledge graph. Run this first to initialize, or let hooks keep it updated automatically.
.claude/skills/tirth8205-build-graph/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -73% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -65% | 0% |
Build or incrementally update the knowledge graph for this repository.
list_graph_stats_tool. If last_updated is null, the graph has never been built.full, call build_or_update_graph_tool(full_rebuild=True). Otherwise call build_or_update_graph_tool() for an incremental update.status (ok, partial or error) and summary.code-review-graph install run an update after each edit and before each commit, so manual builds are rarely needed..code-review-graph/graph.db in the repository root..code-review-graphignore are skipped.get_docs_section_tool(section_name="languages").| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 11,710 | 13,919 | +19% | 1 | 1 | 0% | 290 | 490 | +69% | 0 | 0 | — |
case-02 | fail→fail | 18,643 | 14,219 | -24% | 1 | 1 | 0% | 709 | 466 | -34% | 0 | 0 | — |
case-03 | fail→fail | 27,659 | 10,208 | -63% | 1 | 1 | 0% | 3,253 | 535 | -84% | 0 | 0 | — |
case-04 | fail→fail | 13,770 | 4,320 | -69% | 1 | 1 | 0% | 1,079 | 563 | -48% | 0 | 0 | — |
case-05 | fail→fail | 11,527 | 9,947 | -14% | 1 | 1 | 0% | 705 | 435 | -38% | 0 | 0 | — |
case-06 | fail→pass | 12,480 | 8,011 | -36% | 1 | 1 | 0% | 956 | 524 | -45% | 0 | 0 | — |
case-07 | fail→fail | 32,175 | 7,511 | -77% | 1 | 1 | 0% | 1,014 | 633 | -38% | 0 | 0 | — |
case-08 | fail→fail | 18,307 | 11,042 | -40% | 1 | 1 | 0% | 1,774 | 463 | -74% | 0 | 0 | — |
case-09 | fail→pass | 17,633 | 12,629 | -28% | 1 | 1 | 0% | 2,094 | 635 | -70% | 0 | 0 | — |
case-10 | fail→pass | 23,044 | 17,852 | -23% | 1 | 1 | 0% | 2,502 | 669 | -73% | 0 | 0 | — |
case-11 | fail→pass | 18,601 | 8,131 | -56% | 1 | 1 | 0% | 2,057 | 641 | -69% | 0 | 0 | — |
case-12 | pass→pass | 11,270 | 9,277 | -18% | 1 | 1 | 0% | 928 | 867 | -7% | 0 | 0 | — |
case-13 | pass→fail | 16,464 | 18,182 | +10% | 1 | 1 | 0% | 2,395 | 539 | -77% | 0 | 0 | — |
case-14 | fail→pass | 10,354 | 7,878 | -24% | 1 | 1 | 0% | 1,643 | 573 | -65% | 0 | 0 | — |
case-15 | fail→pass | 33,911 | 7,714 | -77% | 1 | 1 | 0% | 4,503 | 560 | -88% | 0 | 0 | — |
case-16 | pass→pass | 15,895 | 30,484 | +92% | 1 | 1 | 0% | 1,485 | 498 | -66% | 0 | 0 | — |
case-17 | fail→pass | 13,094 | 7,573 | -42% | 1 | 1 | 0% | 1,238 | 450 | -64% | 0 | 0 | — |
case-18 | fail→pass | 26,626 | 27,342 | +3% | 1 | 1 | 0% | 2,732 | 605 | -78% | 0 | 0 | — |
case-19 | fail→pass | 21,867 | 8,920 | -59% | 1 | 1 | 0% | 2,740 | 684 | -75% | 0 | 0 | — |
case-20 | fail→fail | 8,292 | 17,439 | +110% | 1 | 1 | 0% | 1,031 | 603 | -42% | 0 | 0 | — |
case-21 | pass→pass | 30,455 | 31,079 | +2% | 1 | 1 | 0% | 4,137 | 5,324 | +29% | 0 | 0 | — |
case-22 | pass→pass | 17,277 | 11,333 | -34% | 1 | 1 | 0% | 1,930 | 944 | -51% | 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 13 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 +36 percentage points is the difference between those two pass rates over the 13 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/9/2026 | +14% |
Other measured skills in the registry, with their headline benchmark lift.