Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Advisory router — query Claude, Codex, or Gemini for a quick second opinion. Experimental — only Claude is guaranteed available; other models require dev-mcp-setup configuration.
.claude/skills/evolution-foundation-dev-ask/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 20% | 0% |
Derived from oh-my-claudecode (MIT, Yeachan Heo). Adapted for the EvoNexus Engineering Layer.
EXPERIMENTAL. Quick advisory query to a specific LLM (Claude, Codex, Gemini) for a second opinion. Different from dev-ccg which runs all three in parallel — dev-ask is single-shot.
dev-ccgclaude | codex | geminiworkspace/development/research/[C]ask-{topic}-{date}.md| Model | Required setup | |---|---| | Claude | Native — always available | | Codex | OpenAI API key configured via dev-mcp-setup | | Gemini | Google API key configured via dev-mcp-setup |
If a target model isn't configured, the skill warns: > "{model} is not configured. Only Claude is currently available. Configure via dev-mcp-setup or use dev-ccg to compare available models."
markdown## Ask — {target model} ### Question {question} ### Answer (from {model}) {answer} ### Note [Any caveats — model version, response time, confidence, etc.]
dev-ccg (multi-model parallel)dev-mcp-setup (configures non-Claude APIs)@apex-architect (often the consumer of multi-model perspectives)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-12 | fail→pass | 7,354 | 2,404 | -67% | 1 | 1 | 0% | 1,099 | 834 | -24% | 0 | 0 | — |
case-01 | fail→fail | 9,389 | 12,248 | +30% | 1 | 1 | 0% | 1,578 | 2,739 | +74% | 0 | 0 | — |
case-02 | fail→pass | 16,154 | 10,279 | -36% | 1 | 1 | 0% | 2,875 | 2,073 | -28% | 0 | 0 | — |
case-03 | fail→pass | 12,040 | 5,999 | -50% | 1 | 1 | 0% | 1,839 | 1,330 | -28% | 0 | 0 | — |
case-04 | pass→pass | 8,184 | 6,913 | -16% | 1 | 1 | 0% | 1,421 | 1,656 | +17% | 0 | 0 | — |
case-05 | fail→fail | 11,697 | 2,695 | -77% | 1 | 1 | 0% | 2,025 | 685 | -66% | 0 | 0 | — |
case-06 | fail→fail | 8,996 | 6,576 | -27% | 1 | 1 | 0% | 1,466 | 1,525 | +4% | 0 | 0 | — |
case-07 | fail→pass | 18,194 | 11,370 | -38% | 1 | 1 | 0% | 3,041 | 2,260 | -26% | 0 | 0 | — |
case-08 | fail→pass | 17,270 | 18,009 | +4% | 1 | 1 | 0% | 2,677 | 3,211 | +20% | 0 | 0 | — |
case-09 | fail→pass | 11,021 | 9,289 | -16% | 1 | 1 | 0% | 2,014 | 2,133 | +6% | 0 | 0 | — |
case-10 | fail→pass | 10,911 | 8,540 | -22% | 1 | 1 | 0% | 1,819 | 1,786 | -2% | 0 | 0 | — |
case-11 | fail→pass | 6,239 | 11,927 | +91% | 1 | 1 | 0% | 1,030 | 2,478 | +141% | 0 | 0 | — |
case-13 | pass→pass | 14,336 | 8,907 | -38% | 1 | 1 | 0% | 2,304 | 1,761 | -24% | 0 | 0 | — |
case-14 | fail→pass | 16,377 | 12,565 | -23% | 1 | 1 | 0% | 2,496 | 2,552 | +2% | 0 | 0 | — |
case-15 | pass→pass | 8,261 | 5,480 | -34% | 1 | 1 | 0% | 1,322 | 1,364 | +3% | 0 | 0 | — |
case-16 | fail→pass | 7,906 | 2,344 | -70% | 1 | 1 | 0% | 1,267 | 842 | -34% | 0 | 0 | — |
case-17 | fail→pass | 12,786 | 5,521 | -57% | 1 | 1 | 0% | 2,175 | 1,402 | -36% | 0 | 0 | — |
case-18 | fail→pass | 5,611 | 2,674 | -52% | 1 | 1 | 0% | 819 | 797 | -3% | 0 | 0 | — |
case-19 | fail→pass | 12,888 | 12,821 | -1% | 1 | 1 | 0% | 2,034 | 2,355 | +16% | 0 | 0 | — |
case-20 | fail→pass | 9,815 | 8,665 | -12% | 1 | 1 | 0% | 1,632 | 2,028 | +24% | 0 | 0 | — |
case-21 | fail→pass | 8,586 | 1,907 | -78% | 1 | 1 | 0% | 1,171 | 744 | -36% | 0 | 0 | — |
case-22 | fail→pass | 11,807 | 4,364 | -63% | 1 | 1 | 0% | 1,799 | 1,162 | -35% | 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 +73 percentage points is the difference between those two pass rates over the 22 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.