Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Structured data extraction from deep-read papers — produces comparison tables (method, dataset, metrics, results, limitations). Used by systematic-survey and deep-survey.
.claude/skills/yogsoth-ai-extract-data/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✓→✓ | ＝ Same ✓ | 16% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | 63% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 159% | 0% |
| case-23 | ✓→✓ | ＝ Same ✓ | 13% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -40% | 0% |
Pull structured facts from paper full text into comparison tables.
Subagent — spawned via subagent-spawning/spawn-agent skill.
Data extraction requires careful, systematic reading of multiple full papers simultaneously to ensure consistent extraction across all entries. Dedicated context prevents extraction drift.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | spawn-agent | Spawn a customized CC subagent with full MCP tool access. Used by SOPs that declare execution: subagent. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,645 | 3,502 | -38% | 1 | 1 | 0% | 823 | 497 | -40% | 0 | 0 | — |
case-02 | fail→fail | 4,502 | 11,184 | +148% | 1 | 1 | 0% | 686 | 702 | +2% | 0 | 0 | — |
case-03 | fail→fail | 2,253 | 9,440 | +319% | 1 | 1 | 0% | 313 | 630 | +101% | 0 | 0 | — |
case-04 | fail→fail | 9,158 | 9,850 | +8% | 1 | 1 | 0% | 1,253 | 1,569 | +25% | 0 | 0 | — |
case-05 | fail→fail | 19,267 | 10,618 | -45% | 1 | 1 | 0% | 3,286 | 475 | -86% | 0 | 0 | — |
case-06 | fail→fail | 3,779 | 28,939 | +666% | 1 | 1 | 0% | 593 | 4,816 | +712% | 0 | 0 | — |
case-07 | fail→fail | 12,590 | 24,050 | +91% | 1 | 1 | 0% | 2,023 | 4,105 | +103% | 0 | 0 | — |
case-08 | fail→fail | 9,920 | 14,834 | +50% | 1 | 1 | 0% | 1,627 | 2,449 | +51% | 0 | 0 | — |
case-09 | fail→fail | 6,241 | 4,034 | -35% | 1 | 1 | 0% | 1,021 | 646 | -37% | 0 | 0 | — |
case-10 | fail→fail | 4,802 | 4,282 | -11% | 1 | 1 | 0% | 737 | 728 | -1% | 0 | 0 | — |
case-11 | fail→fail | 2,963 | 4,055 | +37% | 1 | 1 | 0% | 404 | 480 | +19% | 0 | 0 | — |
case-12 | fail→fail | 5,338 | 10,094 | +89% | 1 | 1 | 0% | 744 | 867 | +17% | 0 | 0 | — |
case-13 | fail→fail | 7,604 | 5,181 | -32% | 1 | 1 | 0% | 1,268 | 861 | -32% | 0 | 0 | — |
case-14 | fail→fail | 5,000 | 13,240 | +165% | 1 | 1 | 0% | 711 | 658 | -7% | 0 | 0 | — |
case-15 | fail→fail | 5,462 | 15,127 | +177% | 1 | 1 | 0% | 749 | 873 | +17% | 0 | 0 | — |
case-16 | fail→fail | 4,321 | 21,616 | +400% | 1 | 1 | 0% | 597 | 1,861 | +212% | 0 | 0 | — |
case-17 | fail→fail | 4,899 | 8,269 | +69% | 1 | 1 | 0% | 752 | 588 | -22% | 0 | 0 | — |
case-18 | fail→fail | 4,066 | 8,606 | +112% | 1 | 1 | 0% | 569 | 732 | +29% | 0 | 0 | — |
case-19 | fail→fail | 3,502 | 5,863 | +67% | 1 | 1 | 0% | 529 | 355 | -33% | 0 | 0 | — |
case-20 | pass→pass | 5,969 | 6,516 | +9% | 1 | 1 | 0% | 941 | 1,091 | +16% | 0 | 0 | — |
case-21 | pass→pass | 2,764 | 3,807 | +38% | 1 | 1 | 0% | 408 | 667 | +63% | 0 | 0 | — |
case-22 | pass→pass | 2,735 | 4,944 | +81% | 1 | 1 | 0% | 331 | 857 | +159% | 0 | 0 | — |
case-23 | pass→pass | 15,478 | 14,432 | -7% | 1 | 1 | 0% | 2,470 | 2,798 | +13% | 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. 23 cases were attempted, and 14 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 0 percentage points is the difference between those two pass rates over the 14 comparable cases.
Other measured skills in the registry, with their headline benchmark lift.