Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Redirect to claude-model-inference for Messages API streaming, vision, and structured output patterns. Use when looking for the primary Anthropic workflow. Trigger with "anthropic workflow", "claude main workflow".
.claude/skills/jeremylongshore-clade-core-workflow-a/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 84% | 28 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 28% | 0% |
This skill redirects to clade-model-inference which covers streaming, vision, structured output, and all Messages API patterns.
clade-install-auth setupANTHROPIC_API_KEY configuredThis skill has been replaced. The primary Anthropic workflow is the Messages API, covered in full by clade-model-inference.
client.messages.stream()clade-model-inference| Issue | Solution | |-------|----------| | Skill not found | Run clade-model-inference directly |
typescript// Use claude-model-inference for the full Messages API guide import Anthropic from '@claude-ai/sdk'; const client = new Anthropic(); const stream = client.messages.stream({ model: 'claude-sonnet-4-20250514', max_tokens: 1024, messages: [{ role: 'user', content: 'Hello!' }], });
Run clade-model-inference for the complete guide.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 15,175 | 13,899 | -8% | 1 | 1 | 0% | 3,042 | 3,252 | +7% | 0 | 0 | — |
case-02 | fail→fail | 15,946 | 14,458 | -9% | 1 | 1 | 0% | 3,101 | 3,237 | +4% | 0 | 0 | — |
case-03 | fail→fail | 13,685 | 12,521 | -9% | 1 | 1 | 0% | 2,596 | 2,888 | +11% | 0 | 0 | — |
case-04 | fail→pass | 11,915 | 3,820 | -68% | 1 | 1 | 0% | 1,966 | 1,093 | -44% | 0 | 0 | — |
case-05 | fail→pass | 8,425 | 3,697 | -56% | 1 | 1 | 0% | 1,420 | 1,100 | -23% | 0 | 0 | — |
case-06 | fail→pass | 11,377 | 2,466 | -78% | 1 | 1 | 0% | 1,852 | 724 | -61% | 0 | 0 | — |
case-07 | fail→pass | 15,203 | 8,162 | -46% | 1 | 1 | 0% | 2,311 | 1,809 | -22% | 0 | 0 | — |
case-08 | fail→pass | 4,519 | 2,491 | -45% | 1 | 1 | 0% | 615 | 788 | +28% | 0 | 0 | — |
case-09 | fail→pass | 6,006 | 2,510 | -58% | 1 | 1 | 0% | 843 | 796 | -6% | 0 | 0 | — |
case-10 | fail→pass | 9,365 | 3,292 | -65% | 1 | 1 | 0% | 1,530 | 897 | -41% | 0 | 0 | — |
case-11 | fail→pass | 10,433 | 2,271 | -78% | 1 | 1 | 0% | 1,843 | 729 | -60% | 0 | 0 | — |
case-12 | fail→pass | 6,401 | 3,161 | -51% | 1 | 1 | 0% | 1,011 | 946 | -6% | 0 | 0 | — |
case-13 | fail→pass | 10,268 | 3,437 | -67% | 1 | 1 | 0% | 2,102 | 947 | -55% | 0 | 0 | — |
case-14 | fail→pass | 4,032 | 2,011 | -50% | 1 | 1 | 0% | 614 | 668 | +9% | 0 | 0 | — |
case-15 | pass→pass | 3,839 | 1,534 | -60% | 1 | 1 | 0% | 524 | 567 | +8% | 0 | 0 | — |
case-16 | fail→pass | 12,477 | 3,170 | -75% | 1 | 1 | 0% | 1,939 | 878 | -55% | 0 | 0 | — |
case-17 | fail→pass | 6,975 | 2,314 | -67% | 1 | 1 | 0% | 1,212 | 765 | -37% | 0 | 0 | — |
case-18 | fail→pass | 11,489 | 2,505 | -78% | 1 | 1 | 0% | 566 | 745 | +32% | 0 | 0 | — |
case-19 | fail→pass | 5,650 | 1,781 | -68% | 1 | 1 | 0% | 846 | 624 | -26% | 0 | 0 | — |
case-20 | pass→pass | 21,016 | 11,067 | -47% | 1 | 1 | 0% | 1,521 | 1,699 | +12% | 0 | 0 | — |
case-21 | pass→pass | 2,898 | 3,665 | +26% | 1 | 1 | 0% | 410 | 1,044 | +155% | 0 | 0 | — |
case-22 | pass→pass | 4,124 | 3,461 | -16% | 1 | 1 | 0% | 660 | 907 | +37% | 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 21 counted toward the lift figure. The other 1 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 +68 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.