Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create a minimal working Cohere example with Chat, Embed, and Rerank. Use when starting a new Cohere integration, testing your setup, or learning basic Cohere API v2 patterns. Trigger with phrases like "cohere hello world", "cohere example", "cohere quick start", "simple cohere code".
.claude/skills/jeremylongshore-cohere-hello-world/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 58% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 77% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 116% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 26% | 0% |
Prove one selected endpoint end to end while keeping model choice, input type, output bounds, and evidence explicit.
Use Read, Glob, and Grep to inspect code, configuration, and evidence. Use WebFetch only for current Cohere primary documentation. Use Write or Edit only when the user requested implementation and the exact target files are known; never write credentials or customer content.
messages array and returns content under the response message.input_type and embedding_types; use search_document for indexed passages and search_query for queries.rerank-v4.0-pro for quality or rerank-v4.0-fast for measured latency needs.Use an environment-specific key injected from an approved secret manager. Never print, persist, commit, or place CO_API_KEY in an example. Confirm access with the least costly bounded operation appropriate to the task, and treat key creation, rotation, revocation, role changes, and production-capacity requests as owner-approved actions.
cohere-install-auth and choose exactly one endpoint.Do not expose or rotate keys, change Cohere Team roles, accept commercial terms, enable sensitive production data, increase spend or capacity, switch production models, send a support bundle, or execute model-proposed side effects without the accountable owner's approval. Keep diagnosis read-only unless implementation was requested.
Return the resolved API and model contract, files or settings inspected, evidence collected, validation result, remaining risk, owner, and rollback or next action. Redact keys, authorization headers, prompts, retrieved documents, embeddings, customer identifiers, and unrestricted environment output.
| Condition | Response | |---|---| | 400 | Check required v2 fields and endpoint-specific input types. | | 401 | Re-run the read-only authentication probe. | | 404 model | Resolve an accessible live model instead of guessing an alias. | | 429 | Honor the endpoint limit and stop the smoke test. |
Use this compact handoff shape to keep the selected scope, validation evidence, and operational result reviewable.
Input:
textendpoint=embed; input=two-public-sentences; model=resolve-live; max-requests=1
Expected handoff:
textendpoint=embed-v2; model=resolved-id; vectors=2; validation=pass
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 6,631 | 4,045 | -39% | 1 | 1 | 0% | 1,263 | 1,994 | +58% | 0 | 0 | — |
case-02 | fail→pass | 6,072 | 4,566 | -25% | 1 | 1 | 0% | 1,281 | 2,118 | +65% | 0 | 0 | — |
case-03 | pass→pass | 14,276 | 13,512 | -5% | 1 | 1 | 0% | 2,947 | 4,134 | +40% | 0 | 0 | — |
case-04 | pass→pass | 17,811 | 18,441 | +4% | 1 | 1 | 0% | 3,266 | 4,912 | +50% | 0 | 0 | — |
case-05 | pass→pass | 15,273 | 13,691 | -10% | 1 | 1 | 0% | 3,091 | 3,995 | +29% | 0 | 0 | — |
case-06 | fail→pass | 5,607 | 3,662 | -35% | 1 | 1 | 0% | 1,132 | 2,006 | +77% | 0 | 0 | — |
case-07 | fail→pass | 4,485 | 2,577 | -43% | 1 | 1 | 0% | 798 | 1,727 | +116% | 0 | 0 | — |
case-08 | fail→pass | 9,861 | 5,928 | -40% | 1 | 1 | 0% | 1,891 | 2,391 | +26% | 0 | 0 | — |
case-09 | pass→pass | 9,030 | 5,489 | -39% | 1 | 1 | 0% | 1,855 | 2,468 | +33% | 0 | 0 | — |
case-10 | pass→pass | 7,934 | 3,706 | -53% | 1 | 1 | 0% | 1,584 | 1,921 | +21% | 0 | 0 | — |
case-11 | pass→pass | 4,440 | 2,401 | -46% | 1 | 1 | 0% | 767 | 1,600 | +109% | 0 | 0 | — |
case-12 | pass→pass | 8,046 | 3,482 | -57% | 1 | 1 | 0% | 1,576 | 1,803 | +14% | 0 | 0 | — |
case-13 | pass→pass | 5,210 | 3,564 | -32% | 1 | 1 | 0% | 954 | 1,818 | +91% | 0 | 0 | — |
case-14 | pass→pass | 6,968 | 4,116 | -41% | 1 | 1 | 0% | 1,316 | 1,945 | +48% | 0 | 0 | — |
case-15 | pass→pass | 2,568 | 1,867 | -27% | 1 | 1 | 0% | 452 | 1,465 | +224% | 0 | 0 | — |
case-16 | pass→pass | 5,073 | 3,503 | -31% | 1 | 1 | 0% | 1,056 | 1,885 | +79% | 0 | 0 | — |
case-17 | fail→pass | 5,052 | 2,988 | -41% | 1 | 1 | 0% | 921 | 1,736 | +88% | 0 | 0 | — |
case-18 | pass→pass | 10,323 | 4,192 | -59% | 1 | 1 | 0% | 2,073 | 2,041 | -2% | 0 | 0 | — |
case-19 | pass→pass | 5,129 | 2,613 | -49% | 1 | 1 | 0% | 966 | 1,722 | +78% | 0 | 0 | — |
case-20 | pass→pass | 7,018 | 1,677 | -76% | 1 | 1 | 0% | 1,252 | 1,459 | +17% | 0 | 0 | — |
case-21 | pass→pass | 1,666 | 1,739 | +4% | 1 | 1 | 0% | 277 | 1,417 | +412% | 0 | 0 | — |
case-22 | pass→pass | 6,147 | 3,705 | -40% | 1 | 1 | 0% | 1,194 | 1,900 | +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 +27 percentage points is the difference between those two pass rates over the 22 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
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.