Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply Ulrich's Critical Systems Heuristics 12 questions across 4 dimensions (motivation, control, expertise, legitimacy) comparing is vs ought.
.claude/skills/yogsoth-ai-csh-12-question/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-23 | ✓→✗ | ▼ Worse | 77% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | 39% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | 142% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 22% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | 7% | 0% |
Critical Systems Heuristics boundary analysis.
Subagent — spawned via subagent-spawning/spawn-agent.
One unit = one CSH 12-question analysis pass.
<!-- 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 | 20,980 | 21,716 | +4% | 1 | 1 | 0% | 3,348 | 3,577 | +7% | 0 | 0 | — |
case-02 | fail→fail | 21,833 | 24,613 | +13% | 1 | 1 | 0% | 3,440 | 4,174 | +21% | 0 | 0 | — |
case-03 | fail→fail | 21,929 | 22,356 | +2% | 1 | 1 | 0% | 4,029 | 4,250 | +5% | 0 | 0 | — |
case-04 | fail→fail | 22,700 | 24,712 | +9% | 1 | 1 | 0% | 3,649 | 4,143 | +14% | 0 | 0 | — |
case-05 | fail→fail | 20,398 | 18,229 | -11% | 1 | 1 | 0% | 3,016 | 2,912 | -3% | 0 | 0 | — |
case-06 | fail→fail | 18,285 | 20,783 | +14% | 1 | 1 | 0% | 2,959 | 3,197 | +8% | 0 | 0 | — |
case-07 | fail→fail | 18,825 | 19,718 | +5% | 1 | 1 | 0% | 3,146 | 3,117 | -1% | 0 | 0 | — |
case-08 | fail→fail | 21,489 | 33,538 | +56% | 1 | 1 | 0% | 3,306 | 5,128 | +55% | 0 | 0 | — |
case-09 | fail→fail | 20,058 | 29,022 | +45% | 1 | 1 | 0% | 3,221 | 4,838 | +50% | 0 | 0 | — |
case-10 | fail→fail | 20,430 | 26,443 | +29% | 1 | 1 | 0% | 2,746 | 4,133 | +51% | 0 | 0 | — |
case-11 | fail→fail | 21,681 | 23,459 | +8% | 1 | 1 | 0% | 3,736 | 2,976 | -20% | 0 | 0 | — |
case-12 | fail→fail | 20,172 | 17,854 | -11% | 1 | 1 | 0% | 3,391 | 2,739 | -19% | 0 | 0 | — |
case-13 | fail→fail | 19,747 | 26,886 | +36% | 1 | 1 | 0% | 3,052 | 4,286 | +40% | 0 | 0 | — |
case-14 | fail→fail | 19,449 | 14,914 | -23% | 1 | 1 | 0% | 3,232 | 2,438 | -25% | 0 | 0 | — |
case-15 | fail→fail | 20,321 | 23,187 | +14% | 1 | 1 | 0% | 3,167 | 3,489 | +10% | 0 | 0 | — |
case-16 | fail→fail | 22,636 | 24,223 | +7% | 1 | 1 | 0% | 3,321 | 3,856 | +16% | 0 | 0 | — |
case-17 | fail→fail | 20,812 | 19,067 | -8% | 1 | 1 | 0% | 3,267 | 2,971 | -9% | 0 | 0 | — |
case-18 | fail→fail | 22,293 | 19,860 | -11% | 1 | 1 | 0% | 3,503 | 3,008 | -14% | 0 | 0 | — |
case-19 | fail→fail | 48,458 | 38,362 | -21% | 1 | 1 | 0% | 3,770 | 6,109 | +62% | 0 | 0 | — |
case-20 | pass→pass | 12,917 | 17,935 | +39% | 1 | 1 | 0% | 2,055 | 2,860 | +39% | 0 | 0 | — |
case-21 | pass→pass | 11,721 | 28,038 | +139% | 1 | 1 | 0% | 1,793 | 4,342 | +142% | 0 | 0 | — |
case-22 | pass→pass | 13,076 | 15,601 | +19% | 1 | 1 | 0% | 2,181 | 2,652 | +22% | 0 | 0 | — |
case-23 | pass→fail | 22,724 | 41,359 | +82% | 1 | 1 | 0% | 3,573 | 6,324 | +77% | 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. The headline lift of -4 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Other measured skills in the registry, with their headline benchmark lift.