Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Multi-phase project assessment with scoring. Triggered by "/assess" to evaluate project health across dimensions and generate a comprehensive report.
.claude/skills/x-cmd-assess/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 88% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 698% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 85% | 0% |
Use when:
/assessDo NOT use when:
/scan)/check)Generate target.tsv mapping each file to applicable rules.
root target rule-to-check
./ src/auth.sh P06-var-010 P06-var-020Generate result/<rule-id>.tsv for each rule.
target score hint
src/auth.sh 90 variable naming issueScoring: 0-100 (100 = full compliance)
Aggregate into comprehensive report.
markdown## Assessment Summary ### Overall Score: [N]/100 ### Rules Breakdown | Rule | Score | Issues | |------|-------|--------| | ... | ... | ... | ### Top Violations 1. [file:line — issue] ### Recommendations - [Action items]
If no ruleset available, use general heuristics:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 8,963 | 11,907 | +33% | 1 | 1 | 0% | 1,186 | 2,224 | +88% | 0 | 0 | — |
case-02 | fail→fail | 21,285 | 25,131 | +18% | 1 | 1 | 0% | 1,768 | 1,665 | -6% | 0 | 0 | — |
case-03 | fail→fail | 13,713 | 6,837 | -50% | 1 | 1 | 0% | 2,162 | 829 | -62% | 0 | 0 | — |
case-04 | pass→pass | 4,703 | 5,871 | +25% | 1 | 1 | 0% | 647 | 1,191 | +84% | 0 | 0 | — |
case-05 | pass→pass | 8,629 | 6,015 | -30% | 1 | 1 | 0% | 1,519 | 1,236 | -19% | 0 | 0 | — |
case-06 | fail→pass | 6,285 | 5,420 | -14% | 1 | 1 | 0% | 121 | 965 | +698% | 0 | 0 | — |
case-07 | fail→pass | 28,500 | 4,265 | -85% | 1 | 1 | 0% | 2,149 | 816 | -62% | 0 | 0 | — |
case-08 | fail→pass | 14,469 | 5,038 | -65% | 1 | 1 | 0% | 2,250 | 985 | -56% | 0 | 0 | — |
case-09 | pass→pass | 11,747 | 2,536 | -78% | 1 | 1 | 0% | 1,772 | 590 | -67% | 0 | 0 | — |
case-10 | pass→pass | 12,228 | 9,339 | -24% | 1 | 1 | 0% | 1,989 | 1,773 | -11% | 0 | 0 | — |
case-11 | fail→pass | 11,956 | 12,102 | +1% | 1 | 1 | 0% | 1,315 | 2,436 | +85% | 0 | 0 | — |
case-12 | fail→pass | 24,045 | 5,703 | -76% | 1 | 1 | 0% | 1,411 | 928 | -34% | 0 | 0 | — |
case-13 | fail→fail | 30,700 | 3,654 | -88% | 1 | 1 | 0% | 5,306 | 845 | -84% | 0 | 0 | — |
case-14 | fail→pass | 12,357 | 11,661 | -6% | 1 | 1 | 0% | 1,964 | 1,275 | -35% | 0 | 0 | — |
case-15 | fail→pass | 15,337 | 3,769 | -75% | 1 | 1 | 0% | 2,453 | 878 | -64% | 0 | 0 | — |
case-16 | fail→pass | 20,811 | 4,184 | -80% | 1 | 1 | 0% | 2,208 | 983 | -55% | 0 | 0 | — |
case-17 | fail→pass | 14,162 | 4,254 | -70% | 1 | 1 | 0% | 2,092 | 924 | -56% | 0 | 0 | — |
case-18 | fail→pass | 16,123 | 9,234 | -43% | 1 | 1 | 0% | 2,352 | 1,313 | -44% | 0 | 0 | — |
case-19 | fail→pass | 11,212 | 7,191 | -36% | 1 | 1 | 0% | 1,721 | 1,495 | -13% | 0 | 0 | — |
case-20 | fail→pass | 14,041 | 3,914 | -72% | 1 | 1 | 0% | 1,829 | 862 | -53% | 0 | 0 | — |
case-21 | fail→pass | 11,671 | 2,622 | -78% | 1 | 1 | 0% | 1,773 | 732 | -59% | 0 | 0 | — |
case-22 | fail→pass | 18,143 | 2,640 | -85% | 1 | 1 | 0% | 2,567 | 634 | -75% | 0 | 0 | — |
case-23 | pass→pass | 7,958 | 13,405 | +68% | 1 | 1 | 0% | 1,201 | 680 | -43% | 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, and 21 counted toward the lift figure. The other 2 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 +65 percentage points is the difference between those two pass rates over the 21 comparable cases.
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.