From superpowers-plus
Forces quantitative decision matrix evaluation before user questions on 'should I', approaches, or trade-offs. Auto-selects highest-scoring option unless top 2 within 10% and irreversible.
npx claudepluginhub bordenet/superpowers-plus --plugin superpowers-plusThis skill uses the workspace's default tool permissions.
> **Wrong skill?** Design evaluation -> debate. Creative options -> brainstorming. Stuck -> think-twice.
Mandates invoking relevant skills via tools before any response in coding sessions. Covers access, priorities, and adaptations for Claude Code, Copilot CLI, Gemini CLI.
Share bugs, ideas, or general feedback.
Wrong skill? Design evaluation -> debate. Creative options -> brainstorming. Stuck -> think-twice.
Announce at start: "I am using the quantitative-decision-gate skill to evaluate options quantitatively."
# Example: Deciding between extract-method vs inline
echo "Decision: refactor approach"
echo "| Dimension | Wt | Extract | Inline |"
echo "| Impact | 0.35 | 8 | 4 |"
echo "| Effort (inv) | 0.25 | 6 | 9 |"
echo "| Risk (inv) | 0.25 | 7 | 8 |"
echo "| Reversible | 0.15 | 9 | 9 |"
echo "| Weighted | | 7.35 | 6.85 |"
echo "Margin: 7% -> AUTO-SELECT extract-method"
State: "I need to decide: [decision]". List options (min 2, max 5).
| Dimension | Weight | Option A | Option B |
|---|---|---|---|
| Impact | 0.35 | 8 | 6 |
| Effort (inv) | 0.25 | 9 (low) | 4 (high) |
| Risk (inv) | 0.25 | 7 (low) | 8 (low) |
| Reversibility | 0.15 | 9 | 5 |
| Weighted | 8.15 | 5.85 |
Score 1-10 (10=best). Effort is INVERSE (high score = low effort).
Record: decision, scores, 1-sentence rationale.
| Anti-Pattern | Detection | Correction |
|---|---|---|
| Asking user a quantifiable decision | "should I" + options exist | Build matrix, auto-select |
| Victory-claiming without data | "better" without numbers | Produce the numbers |
| False ceiling diagnosis | "impossible because X" | Try 2 more approaches first |
| Single-option framing | "Should I do X?" only | Generate >=1 alternative |
| Gut-feel decision | "feels right" | Score it. Numbers confirm or deny |
| Failure | Detection | Recovery |
|---|---|---|
| Matrix not followed | Decision != highest score | Re-examine scoring |
| All options tie | Within 5% | Add context-specific 5th dimension |
| Scoring bias | One option all 9s/10s | Red-team: argue FOR the loser |
| Decision not acted on | TODO but no action | Execute or defer with reason |