From superpowers-plus
Enforces cross-validation with at least two methods, completeness verification, and confidence qualification before reporting any metric, percentage, or count like coverage or pass rates.
npx claudepluginhub bordenet/superpowers-plus --plugin superpowers-plusThis skill uses the workspace's default tool permissions.
> **Wrong skill?** Verifying work -> verification-before-completion. Completeness audit -> exhaustive-audit-validation.
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? Verifying work -> verification-before-completion. Completeness audit -> exhaustive-audit-validation.
Announce at start: "I am using the measurement-integrity skill to validate this metric."
# Example: Cross-validating skill count
method1=$(find skills -name "skill.md" | wc -l)
method2=$(find skills -name "skill.md" -not -path "*/_*" | wc -l)
echo "Method 1 (all): $method1"
echo "Method 2 (excl support): $method2"
echo "Delta: $((method1 - method2)) support files"
# Report: "72 skills (filesystem-verified, 2 methods)"
Verify universe before reporting percentages. Example: total= 0 WRONG: "65/65" without verifying 65 is correct.
Report: Metric + Method + Universe + Confidence + Caveat
Same methodology? Same universe? Same test suite?
| Anti-Pattern | Example | Fix |
|---|---|---|
| % without universe | "95% coverage" | "95% (62/65 skills)" |
| Changed methodology | 14/14 -> 14/17 | Disclose: "expanded suite" |
| Ceiling from one test | "Cannot do better" | Try 3 fixes first |
| Precision theater | "93.846%" | Round: "94%" |
| Failure | Detection | Recovery |
|---|---|---|
| No cross-validation | Missing Method 2 | Add second method |
| Universe changed | Total differs | Acknowledge + rebaseline |
| False precision | Excess decimals | Round to significant digits |
| Stops investigation | "ceiling" claim | Try 2 more approaches |