Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Map all affected parties using CSH 12-question framework, identify jobs-to-be-done, classify by salience. Reveals whose perspective is systematically excluded.
.claude/skills/yogsoth-ai-stakeholder-mapping/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-02 | ✓→✗ | ▼ Worse | -67% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -58% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -72% | 0% |
Map who is affected by a gap and whose perspective is missing.
Need to understand the human/institutional landscape around a research gap — who benefits, who suffers, who is excluded.
| Base SOP | Target | ±10% Range | |----------|--------|------------| | web-search | 40 | 36–44 | | web-research | 15 | 13–17 | | paper-overview | 30 | 27–33 | | paper-search | 20 | 18–22 | | paper-research | 5 | 4–6 |
<HARD-GATE>
| SOP | Done | Target | % |
|-----|------|--------|---|
| web-search | ? | 40 | ? |
| web-research | ? | 15 | ? |
| paper-overview | ? | 30 | ? |
| paper-search | ? | 20 | ? |
| paper-research | ? | 5 | ? |
Budget Gate: OPEN/CLOSED (>=80% required to exit)
</HARD-GATE>Import: web-search, web-research, paper-overview, paper-search, paper-research Subagent: csh-12-question, jtbd-mapping, salience-classification Shared: multi-stakeholder-simulation
Map all affected parties using CSH 12-question framework, identify their jobs-to-be-done, classify by salience (power/legitimacy/urgency). Reveal whose perspective is systematically excluded.
Stakeholder Map — CSH matrix, JTBD per stakeholder, salience classification, excluded perspectives.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | boundary-unfolding | Systematically expose hidden system boundaries — CSH 12-question is/ought comparison, identify excluded stakeholders, reveal blind spots. Combines csh-12-question, jtbd-mapping, and salience-classification SOPs. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | csh-12-question | Apply Ulrich's Critical Systems Heuristics 12 questions across 4 dimensions (motivation, control, expertise, legitimacy) comparing is vs ought. | | deep-insight-multi-stakeholder-simulation | Simulate multiple stakeholder perspectives evaluating a research gap, method, or proposal. Identifies blind spots from single-perspective analysis. | | jtbd-mapping | Map stakeholder Jobs-to-be-Done — functional, emotional, and social jobs for each affected party. Identifies unserved jobs as opportunity signals. | | salience-classification | Classify stakeholders by Mitchell et al. framework (Power, Legitimacy, Urgency). Assigns salience category and identifies systematically excluded parties. |
<!-- 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 | 51,338 | 23,563 | -54% | 1 | 1 | 0% | 7,324 | 1,667 | -77% | 0 | 0 | — |
case-02 | pass→fail | 34,474 | 29,195 | -15% | 1 | 1 | 0% | 4,683 | 1,536 | -67% | 0 | 0 | — |
case-03 | fail→fail | 28,984 | 21,338 | -26% | 1 | 1 | 0% | 4,622 | 1,527 | -67% | 0 | 0 | — |
case-04 | pass→fail | 28,007 | 12,008 | -57% | 1 | 1 | 0% | 3,837 | 1,623 | -58% | 0 | 0 | — |
case-05 | pass→fail | 49,544 | 18,126 | -63% | 1 | 1 | 0% | 6,468 | 1,834 | -72% | 0 | 0 | — |
case-06 | fail→pass | 54,851 | 59,736 | +9% | 1 | 1 | 0% | 6,067 | 8,948 | +47% | 0 | 0 | — |
case-07 | fail→fail | 30,137 | 23,092 | -23% | 1 | 1 | 0% | 4,323 | 1,542 | -64% | 0 | 0 | — |
case-21 | pass→fail | 44,307 | 62,174 | +40% | 1 | 1 | 0% | 8,234 | 8,959 | +9% | 0 | 0 | — |
case-08 | pass→fail | 51,036 | 16,310 | -68% | 1 | 1 | 0% | 7,115 | 1,438 | -80% | 0 | 0 | — |
case-09 | fail→fail | 27,729 | 17,874 | -36% | 1 | 1 | 0% | 3,206 | 2,232 | -30% | 0 | 0 | — |
case-10 | fail→fail | 48,116 | 22,881 | -52% | 1 | 1 | 0% | 6,761 | 1,508 | -78% | 0 | 0 | — |
case-11 | pass→fail | 32,618 | 44,433 | +36% | 1 | 1 | 0% | 4,247 | 1,590 | -63% | 0 | 0 | — |
case-12 | fail→fail | 29,589 | 18,764 | -37% | 1 | 1 | 0% | 4,004 | 1,590 | -60% | 0 | 0 | — |
case-13 | pass→fail | 24,563 | 20,667 | -16% | 1 | 1 | 0% | 2,899 | 2,184 | -25% | 0 | 0 | — |
case-14 | pass→fail | 20,395 | 20,050 | -2% | 1 | 1 | 0% | 3,026 | 1,651 | -45% | 0 | 0 | — |
case-15 | pass→fail | 24,935 | 15,254 | -39% | 1 | 1 | 0% | 3,004 | 2,045 | -32% | 0 | 0 | — |
case-16 | pass→fail | 21,869 | 27,518 | +26% | 1 | 1 | 0% | 3,119 | 1,997 | -36% | 0 | 0 | — |
case-17 | pass→fail | 32,783 | 21,735 | -34% | 1 | 1 | 0% | 3,878 | 1,786 | -54% | 0 | 0 | — |
case-18 | fail→fail | 52,686 | 19,512 | -63% | 1 | 1 | 0% | 8,219 | 1,481 | -82% | 0 | 0 | — |
case-19 | fail→pass | 38,023 | 52,477 | +38% | 1 | 1 | 0% | 5,705 | 8,943 | +57% | 0 | 0 | — |
case-20 | pass→pass | 29,530 | 49,845 | +69% | 1 | 1 | 0% | 5,698 | 8,975 | +58% | 0 | 0 | — |
case-22 | pass→fail | 123,130 | 54,641 | -56% | 1 | 1 | 0% | 8,242 | 8,967 | +9% | 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 8 counted toward the lift figure. The other 14 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 -45 percentage points is the difference between those two pass rates over the 8 comparable cases. 17 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.