Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Identify bottleneck dimensions from radar data with severity ranking.
.claude/skills/yogsoth-ai-convergence-bottleneck-identification/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 278% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 332% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 97% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 95% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 251% | 0% |
Analyze radar chart data to identify which dimensions are bottlenecks — the limiting factors that constrain overall feasibility. Produces a severity-ranked list of bottlenecks with analysis of why each is limiting.
Spawns a subagent that:
Bottleneck identification requires comparative analysis across dimensions and judgment about which low scores are truly limiting vs. acceptable given the context.
Output MUST include: at least 1 bottleneck identified, severity ranking, and prioritized recommendation. Reject if no bottlenecks are identified (even high-readiness candidates have relative bottlenecks).
<!-- 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 | pass→fail | 31,080 | 25,586 | -18% | 1 | 1 | 0% | 6,216 | 5,845 | -6% | 0 | 0 | — |
case-02 | fail→fail | 3,168 | 20,342 | +542% | 1 | 1 | 0% | 500 | 3,536 | +607% | 0 | 0 | — |
case-03 | pass→pass | 13,583 | 24,633 | +81% | 1 | 1 | 0% | 2,804 | 5,692 | +103% | 0 | 0 | — |
case-04 | fail→pass | 3,856 | 12,586 | +226% | 1 | 1 | 0% | 714 | 2,701 | +278% | 0 | 0 | — |
case-05 | fail→fail | 13,747 | 10,461 | -24% | 1 | 1 | 0% | 2,347 | 954 | -59% | 0 | 0 | — |
case-06 | pass→pass | 8,821 | 13,181 | +49% | 1 | 1 | 0% | 1,423 | 2,602 | +83% | 0 | 0 | — |
case-07 | fail→fail | 18,241 | 14,135 | -23% | 1 | 1 | 0% | 3,205 | 2,687 | -16% | 0 | 0 | — |
case-08 | fail→pass | 3,046 | 12,407 | +307% | 1 | 1 | 0% | 511 | 2,205 | +332% | 0 | 0 | — |
case-09 | fail→pass | 9,601 | 16,469 | +72% | 1 | 1 | 0% | 1,611 | 3,177 | +97% | 0 | 0 | — |
case-10 | fail→pass | 7,570 | 14,493 | +91% | 1 | 1 | 0% | 1,369 | 2,663 | +95% | 0 | 0 | — |
case-11 | fail→pass | 3,356 | 10,932 | +226% | 1 | 1 | 0% | 589 | 2,066 | +251% | 0 | 0 | — |
case-12 | fail→pass | 8,780 | 15,229 | +73% | 1 | 1 | 0% | 1,468 | 2,900 | +98% | 0 | 0 | — |
case-13 | pass→pass | 13,239 | 17,370 | +31% | 1 | 1 | 0% | 1,934 | 2,996 | +55% | 0 | 0 | — |
case-14 | fail→fail | 4,602 | 11,626 | +153% | 1 | 1 | 0% | 771 | 2,241 | +191% | 0 | 0 | — |
case-15 | pass→pass | 10,949 | 17,475 | +60% | 1 | 1 | 0% | 1,736 | 3,287 | +89% | 0 | 0 | — |
case-16 | fail→pass | 7,637 | 12,847 | +68% | 1 | 1 | 0% | 1,210 | 2,308 | +91% | 0 | 0 | — |
case-17 | fail→pass | 2,623 | 12,828 | +389% | 1 | 1 | 0% | 432 | 2,445 | +466% | 0 | 0 | — |
case-18 | fail→pass | 5,631 | 13,709 | +143% | 1 | 1 | 0% | 1,126 | 2,547 | +126% | 0 | 0 | — |
case-19 | fail→pass | 7,011 | 11,514 | +64% | 1 | 1 | 0% | 1,102 | 2,198 | +99% | 0 | 0 | — |
case-20 | fail→pass | 10,078 | 22,468 | +123% | 1 | 1 | 0% | 1,563 | 3,011 | +93% | 0 | 0 | — |
case-21 | fail→pass | 17,874 | 15,102 | -16% | 1 | 1 | 0% | 2,570 | 2,559 | -0% | 0 | 0 | — |
case-22 | fail→fail | 8,875 | 9,404 | +6% | 1 | 1 | 0% | 1,520 | 1,817 | +20% | 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 21 counted toward the lift figure. The other 1 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 +50 percentage points is the difference between those two pass rates over the 21 comparable cases. 1 case got worse with the skill loaded, and it is 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.