Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when the user types /agentlas-local, /agentlas local, or /hep-local to staff only from registered Local packages.
.claude/skills/agentlas-ai-agentlas-local/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✓→✗ | ▼ Worse | 167% | 0% |
| case-06 | ✓→✗ | ▼ Worse | -26% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 9% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | 50% | 0% |
| case-02 | ✗→✗ | ＝ Same ✗ | -85% | 0% |
Identical to /skill:hep-local and /agentlas local <request>. Locate the sibling skill directory hep-local (i.e. ../hep-local/SKILL.md relative to this file, under the same kimi/skills/ root this skill was loaded from), read its SKILL.md, and follow its instructions exactly — treating everything typed after /skill:agentlas-local as that command's request.
Do not improvise a separate workflow and do not summarize hep-local/SKILL.md from memory; that file is the sole authority for this command's behavior.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,955 | 7,644 | +54% | 1 | 1 | 0% | 335 | 502 | +50% | 0 | 0 | — |
case-02 | fail→fail | 26,357 | 12,608 | -52% | 1 | 1 | 0% | 3,173 | 474 | -85% | 0 | 0 | — |
case-03 | fail→fail | 12,021 | 9,878 | -18% | 1 | 1 | 0% | 1,906 | 593 | -69% | 0 | 0 | — |
case-04 | pass→pass | 20,661 | 49,656 | +140% | 1 | 1 | 0% | 2,760 | 3,011 | +9% | 0 | 0 | — |
case-05 | pass→fail | 9,108 | 35,957 | +295% | 1 | 1 | 0% | 1,531 | 4,086 | +167% | 0 | 0 | — |
case-06 | pass→fail | 5,380 | 7,860 | +46% | 1 | 1 | 0% | 692 | 509 | -26% | 0 | 0 | — |
case-07 | fail→fail | 24,950 | 10,420 | -58% | 1 | 1 | 0% | 173 | 533 | +208% | 0 | 0 | — |
case-08 | fail→fail | 42,538 | 76,914 | +81% | 1 | 1 | 0% | 2,433 | 2,551 | +5% | 0 | 0 | — |
case-09 | fail→fail | 22,156 | 10,601 | -52% | 1 | 1 | 0% | 2,679 | 531 | -80% | 0 | 0 | — |
case-10 | fail→fail | 8,212 | 9,366 | +14% | 1 | 1 | 0% | 285 | 474 | +66% | 0 | 0 | — |
case-11 | fail→fail | 6,012 | 6,255 | +4% | 1 | 1 | 0% | 305 | 377 | +24% | 0 | 0 | — |
case-12 | fail→fail | 18,947 | 8,386 | -56% | 1 | 1 | 0% | 2,446 | 453 | -81% | 0 | 0 | — |
case-13 | fail→fail | 9,818 | 11,092 | +13% | 1 | 1 | 0% | 316 | 830 | +163% | 0 | 0 | — |
case-14 | fail→fail | 4,069 | 7,314 | +80% | 1 | 1 | 0% | 257 | 515 | +100% | 0 | 0 | — |
case-15 | fail→fail | 7,166 | 15,150 | +111% | 1 | 1 | 0% | 262 | 546 | +108% | 0 | 0 | — |
case-16 | fail→fail | 11,826 | 12,086 | +2% | 1 | 1 | 0% | 1,805 | 1,074 | -40% | 0 | 0 | — |
case-17 | fail→fail | 17,750 | 14,051 | -21% | 1 | 1 | 0% | 3,046 | 630 | -79% | 0 | 0 | — |
case-18 | fail→fail | 16,732 | 8,104 | -52% | 1 | 1 | 0% | 2,916 | 381 | -87% | 0 | 0 | — |
case-19 | fail→fail | 39,944 | 8,990 | -77% | 1 | 1 | 0% | 1,168 | 400 | -66% | 0 | 0 | — |
case-20 | fail→fail | 8,788 | 18,276 | +108% | 1 | 1 | 0% | 284 | 1,358 | +378% | 0 | 0 | — |
case-21 | fail→fail | 4,818 | 8,774 | +82% | 1 | 1 | 0% | 375 | 603 | +61% | 0 | 0 | — |
case-22 | fail→fail | 3,733 | 7,423 | +99% | 1 | 1 | 0% | 293 | 465 | +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, and 3 counted toward the lift figure. The other 19 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 -50 percentage points is the difference between those two pass rates over the 3 comparable cases. 13 cases got worse with the skill loaded, and they are included in that figure.
Other measured skills in the registry, with their headline benchmark lift.