Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Classification guidelines for entries in temp_entries.md. Each entry has a title containing the markdown file name and section name. USE FOR: Classifying new entries in temp_entries.md into *.md section files. DO NOT USE FOR: Adding entries to temp_entries.md or moving entries between sections.
.claude/skills/kimtth-add-temp-entries-to-sections/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 73% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -56% | 0% |
temp_entries.md is the staging file where new entries from temp.md are formatted before being inserted into the section files. Each entry in temp_entries.md should have a title indicating the target markdown file and section name for clarity.
Steps for classification:
section/ it belongs to and use the exact current heading from that file's Contents block or heading text.temp_entries.md to the appropriate section files (azure.md, applications.md, models_research.md, best_practices.md, tools_extra.md) under the correct live section headings.temp_entries.md to keep it clean and track which entries have been processed.Do not classify hand-curated entries into generated index files such as section/x_llm_apps.md, section/x_llm_papers.md, or section/x_popular_papers.md. Use the generator-specific skills for those files.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,341 | 4,662 | +40% | 1 | 1 | 0% | 331 | 591 | +79% | 0 | 0 | — |
case-02 | fail→fail | 2,863 | 4,415 | +54% | 1 | 1 | 0% | 272 | 515 | +89% | 0 | 0 | — |
case-03 | fail→fail | 4,674 | 4,670 | -0% | 1 | 1 | 0% | 220 | 523 | +138% | 0 | 0 | — |
case-04 | fail→fail | 5,945 | 4,589 | -23% | 1 | 1 | 0% | 268 | 626 | +134% | 0 | 0 | — |
case-05 | fail→fail | 4,175 | 2,276 | -45% | 1 | 1 | 0% | 664 | 740 | +11% | 0 | 0 | — |
case-06 | fail→fail | 4,983 | 2,388 | -52% | 1 | 1 | 0% | 868 | 740 | -15% | 0 | 0 | — |
case-07 | fail→fail | 9,230 | 6,240 | -32% | 1 | 1 | 0% | 1,145 | 603 | -47% | 0 | 0 | — |
case-08 | fail→fail | 10,562 | 5,349 | -49% | 1 | 1 | 0% | 1,485 | 539 | -64% | 0 | 0 | — |
case-09 | fail→pass | 5,935 | 3,413 | -42% | 1 | 1 | 0% | 876 | 953 | +9% | 0 | 0 | — |
case-10 | fail→fail | 3,512 | 5,008 | +43% | 1 | 1 | 0% | 517 | 616 | +19% | 0 | 0 | — |
case-11 | fail→fail | 33,179 | 4,225 | -87% | 1 | 1 | 0% | 341 | 564 | +65% | 0 | 0 | — |
case-12 | fail→pass | 4,850 | 1,682 | -65% | 1 | 1 | 0% | 693 | 602 | -13% | 0 | 0 | — |
case-13 | fail→fail | 5,485 | 4,970 | -9% | 1 | 1 | 0% | 386 | 525 | +36% | 0 | 0 | — |
case-14 | pass→pass | 6,088 | 3,932 | -35% | 1 | 1 | 0% | 1,030 | 1,082 | +5% | 0 | 0 | — |
case-15 | fail→pass | 3,479 | 2,579 | -26% | 1 | 1 | 0% | 533 | 725 | +36% | 0 | 0 | — |
case-16 | fail→fail | 2,950 | 5,182 | +76% | 1 | 1 | 0% | 320 | 567 | +77% | 0 | 0 | — |
case-17 | fail→pass | 4,182 | 5,121 | +22% | 1 | 1 | 0% | 769 | 1,330 | +73% | 0 | 0 | — |
case-18 | fail→pass | 9,441 | 2,132 | -77% | 1 | 1 | 0% | 1,505 | 669 | -56% | 0 | 0 | — |
case-19 | pass→pass | 6,385 | 4,448 | -30% | 1 | 1 | 0% | 1,108 | 1,034 | -7% | 0 | 0 | — |
case-20 | fail→fail | 1,895 | 3,324 | +75% | 1 | 1 | 0% | 287 | 461 | +61% | 0 | 0 | — |
case-21 | pass→pass | 10,328 | 5,112 | -51% | 1 | 1 | 0% | 1,867 | 1,181 | -37% | 0 | 0 | — |
case-22 | fail→fail | 3,038 | 2,967 | -2% | 1 | 1 | 0% | 538 | 772 | +43% | 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 11 counted toward the lift figure. The other 11 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 +23 percentage points is the difference between those two pass rates over the 11 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.