Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Record tool execution in AxonFlow audit trail after tool calls complete. Use after any tool call for compliance evidence.
.claude/skills/hashgraph-online-post-execute-audit/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -61% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -27% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | -45% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | 28% | 0% |
After tool calls complete, call the audit_tool_call MCP tool with:
tool_name: the tool that was called (e.g., "Bash", "Write")input: the tool input (command, file content, etc.)success: whether the tool call succeededoutput: summary of the result (truncate if large)This creates a compliance-grade audit record with timestamp, policy context, and execution details.
Note: For Bash tool calls, this is handled automatically by the PostToolUse hook. Use this skill for other tool types that are not hook-governed.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 11,100 | 36,825 | +232% | 1 | 1 | 0% | 892 | 1,144 | +28% | 0 | 0 | — |
case-02 | fail→fail | 15,778 | 32,635 | +107% | 1 | 1 | 0% | 3,076 | 1,034 | -66% | 0 | 0 | — |
case-03 | fail→fail | 8,289 | 27,530 | +232% | 1 | 1 | 0% | 430 | 2,789 | +549% | 0 | 0 | — |
case-04 | pass→pass | 8,876 | 7,239 | -18% | 1 | 1 | 0% | 1,339 | 521 | -61% | 0 | 0 | — |
case-05 | pass→pass | 14,120 | 9,590 | -32% | 1 | 1 | 0% | 2,314 | 1,695 | -27% | 0 | 0 | — |
case-06 | pass→pass | 20,814 | 16,081 | -23% | 1 | 1 | 0% | 2,971 | 1,632 | -45% | 0 | 0 | — |
case-07 | fail→pass | 4,994 | 8,962 | +79% | 1 | 1 | 0% | 589 | 560 | -5% | 0 | 0 | — |
case-08 | fail→fail | 11,141 | 24,651 | +121% | 1 | 1 | 0% | 904 | 3,290 | +264% | 0 | 0 | — |
case-09 | fail→fail | 14,453 | 18,446 | +28% | 1 | 1 | 0% | 1,817 | 690 | -62% | 0 | 0 | — |
case-10 | fail→fail | 11,680 | 19,124 | +64% | 1 | 1 | 0% | 1,228 | 867 | -29% | 0 | 0 | — |
case-11 | fail→fail | 4,407 | 12,629 | +187% | 1 | 1 | 0% | 720 | 627 | -13% | 0 | 0 | — |
case-12 | fail→fail | 9,736 | 15,451 | +59% | 1 | 1 | 0% | 848 | 810 | -4% | 0 | 0 | — |
case-13 | fail→fail | 13,682 | 16,343 | +19% | 1 | 1 | 0% | 565 | 967 | +71% | 0 | 0 | — |
case-14 | fail→fail | 13,945 | 24,845 | +78% | 1 | 1 | 0% | 2,490 | 774 | -69% | 0 | 0 | — |
case-15 | fail→fail | 13,423 | 11,058 | -18% | 1 | 1 | 0% | 1,371 | 1,070 | -22% | 0 | 0 | — |
case-16 | fail→fail | 125,664 | 21,497 | -83% | 1 | 1 | 0% | 1,951 | 1,447 | -26% | 0 | 0 | — |
case-17 | fail→fail | 12,490 | 13,173 | +5% | 1 | 1 | 0% | 1,108 | 1,320 | +19% | 0 | 0 | — |
case-18 | fail→fail | 12,360 | 30,356 | +146% | 1 | 1 | 0% | 1,050 | 2,874 | +174% | 0 | 0 | — |
case-19 | fail→fail | 19,077 | 30,759 | +61% | 1 | 1 | 0% | 2,167 | 941 | -57% | 0 | 0 | — |
case-20 | fail→fail | 12,304 | 20,153 | +64% | 1 | 1 | 0% | 1,279 | 1,030 | -19% | 0 | 0 | — |
case-21 | fail→fail | 28,937 | 22,237 | -23% | 1 | 1 | 0% | 1,660 | 749 | -55% | 0 | 0 | — |
case-22 | fail→fail | 6,707 | 28,857 | +330% | 1 | 1 | 0% | 1,008 | 1,046 | +4% | 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 5 counted toward the lift figure. The other 17 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 +5 percentage points is the difference between those two pass rates over the 5 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.