Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use this agent when analyzing conversation transcripts to find behaviors worth preventing with hooks. Triggered by /hookify without arguments.
.claude/skills/kunanonj-agent-conversation-analyzer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 73% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 111% | 0% |
You analyze conversation history to identify problematic Claude Code behaviors that should be prevented with hooks.
git checkout -- file or git restore file after Claude's editFor each identified behavior:
yamlbehavior: "Description of what Claude did wrong" frequency: "How often it occurred" severity: high|medium|low suggested_rule: name: "descriptive-rule-name" event: bash|file|stop|prompt pattern: "regex pattern to match" action: block|warn message: "What to show when triggered"
Prioritize high-frequency, high-severity behaviors first.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 4,946 | 5,844 | +18% | 1 | 1 | 0% | 1,000 | 1,728 | +73% | 0 | 0 | — |
case-02 | fail→fail | 5,849 | 2,277 | -61% | 1 | 1 | 0% | 1,180 | 967 | -18% | 0 | 0 | — |
case-03 | fail→pass | 7,748 | 2,933 | -62% | 1 | 1 | 0% | 1,633 | 1,185 | -27% | 0 | 0 | — |
case-04 | pass→pass | 10,439 | 7,770 | -26% | 1 | 1 | 0% | 2,381 | 2,230 | -6% | 0 | 0 | — |
case-13 | fail→pass | 4,014 | 2,916 | -27% | 1 | 1 | 0% | 822 | 1,155 | +41% | 0 | 0 | — |
case-05 | pass→fail | 4,683 | 4,311 | -8% | 1 | 1 | 0% | 929 | 1,305 | +40% | 0 | 0 | — |
case-06 | pass→pass | 9,991 | 6,068 | -39% | 1 | 1 | 0% | 2,141 | 1,872 | -13% | 0 | 0 | — |
case-07 | fail→pass | 5,537 | 3,379 | -39% | 1 | 1 | 0% | 1,124 | 1,209 | +8% | 0 | 0 | — |
case-08 | fail→pass | 3,083 | 4,178 | +36% | 1 | 1 | 0% | 683 | 1,442 | +111% | 0 | 0 | — |
case-09 | pass→pass | 5,640 | 3,963 | -30% | 1 | 1 | 0% | 1,310 | 1,456 | +11% | 0 | 0 | — |
case-10 | fail→pass | 12,042 | 4,328 | -64% | 1 | 1 | 0% | 2,507 | 1,283 | -49% | 0 | 0 | — |
case-11 | pass→pass | 12,387 | 3,499 | -72% | 1 | 1 | 0% | 2,545 | 1,135 | -55% | 0 | 0 | — |
case-12 | fail→pass | 10,709 | 3,892 | -64% | 1 | 1 | 0% | 2,161 | 1,280 | -41% | 0 | 0 | — |
case-14 | fail→pass | 9,068 | 3,112 | -66% | 1 | 1 | 0% | 2,006 | 1,236 | -38% | 0 | 0 | — |
case-15 | fail→pass | 4,534 | 2,703 | -40% | 1 | 1 | 0% | 824 | 1,040 | +26% | 0 | 0 | — |
case-16 | pass→pass | 2,637 | 4,140 | +57% | 1 | 1 | 0% | 659 | 1,482 | +125% | 0 | 0 | — |
case-17 | pass→pass | 5,470 | 3,059 | -44% | 1 | 1 | 0% | 1,164 | 1,085 | -7% | 0 | 0 | — |
case-18 | fail→pass | 12,520 | 4,101 | -67% | 1 | 1 | 0% | 2,466 | 1,380 | -44% | 0 | 0 | — |
case-19 | fail→pass | 10,048 | 4,091 | -59% | 1 | 1 | 0% | 2,216 | 1,411 | -36% | 0 | 0 | — |
case-20 | pass→pass | 12,942 | 3,100 | -76% | 1 | 1 | 0% | 2,603 | 1,128 | -57% | 0 | 0 | — |
case-21 | fail→pass | 5,065 | 3,169 | -37% | 1 | 1 | 0% | 997 | 1,145 | +15% | 0 | 0 | — |
case-22 | fail→pass | 13,577 | 3,019 | -78% | 1 | 1 | 0% | 2,718 | 1,116 | -59% | 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. The headline lift of +55 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.