Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Collect Cohere debug evidence for support tickets and troubleshooting. Use when encountering persistent issues, preparing support tickets, or collecting diagnostic information for Cohere API problems. Trigger with phrases like "cohere debug", "cohere support bundle", "collect cohere logs", "cohere diagnostic".
.claude/skills/jeremylongshore-cohere-debug-bundle/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 70% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 63% | 0% |
Produce evidence that support can act on without copying credentials, prompts, customer data, or an unrestricted environment dump.
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.
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.
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 | |---|---| | Secret detected | Stop packaging, remove the value, and rotate if exposure occurred. | | Customer text present | Replace it with a synthetic reproducer or approved hash. | | Oversized bundle | Drop unrelated logs and narrow the time window. | | No reproduction | Document observed evidence and mark the failure intermittent. |
Use this compact handoff shape to keep the selected scope, validation evidence, and operational result reviewable.
Input:
textincident=INC-42; window=10m; reproduction=synthetic; destination=local-only
Expected handoff:
textbundle=ready; secrets=0; customer-content=0; checksum=recorded
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 24,194 | 23,355 | -3% | 1 | 1 | 0% | 5,503 | 6,808 | +24% | 0 | 0 | — |
case-02 | fail→pass | 18,280 | 18,338 | +0% | 1 | 1 | 0% | 3,828 | 5,812 | +52% | 0 | 0 | — |
case-03 | fail→pass | 21,052 | 15,584 | -26% | 1 | 1 | 0% | 4,543 | 5,299 | +17% | 0 | 0 | — |
case-04 | pass→pass | 15,576 | 12,484 | -20% | 1 | 1 | 0% | 3,290 | 4,601 | +40% | 0 | 0 | — |
case-05 | pass→pass | 9,966 | 10,195 | +2% | 1 | 1 | 0% | 1,747 | 3,679 | +111% | 0 | 0 | — |
case-06 | pass→pass | 14,138 | 12,418 | -12% | 1 | 1 | 0% | 2,943 | 4,253 | +45% | 0 | 0 | — |
case-07 | pass→pass | 10,996 | 4,604 | -58% | 1 | 1 | 0% | 2,100 | 2,709 | +29% | 0 | 0 | — |
case-08 | fail→pass | 7,108 | 2,767 | -61% | 1 | 1 | 0% | 1,393 | 2,365 | +70% | 0 | 0 | — |
case-09 | pass→pass | 7,309 | 2,067 | -72% | 1 | 1 | 0% | 1,437 | 2,235 | +56% | 0 | 0 | — |
case-10 | pass→pass | 8,933 | 5,381 | -40% | 1 | 1 | 0% | 1,728 | 2,773 | +60% | 0 | 0 | — |
case-11 | pass→pass | 6,940 | 5,826 | -16% | 1 | 1 | 0% | 1,417 | 2,883 | +103% | 0 | 0 | — |
case-12 | pass→pass | 6,884 | 2,775 | -60% | 1 | 1 | 0% | 1,234 | 2,281 | +85% | 0 | 0 | — |
case-13 | fail→pass | 7,330 | 3,431 | -53% | 1 | 1 | 0% | 1,588 | 2,592 | +63% | 0 | 0 | — |
case-14 | pass→pass | 11,501 | 9,388 | -18% | 1 | 1 | 0% | 2,504 | 3,776 | +51% | 0 | 0 | — |
case-15 | fail→pass | 24,538 | 7,190 | -71% | 1 | 1 | 0% | 1,239 | 3,216 | +160% | 0 | 0 | — |
case-16 | fail→pass | 7,685 | 3,114 | -59% | 1 | 1 | 0% | 1,318 | 2,393 | +82% | 0 | 0 | — |
case-17 | pass→pass | 15,588 | 11,725 | -25% | 1 | 1 | 0% | 2,976 | 4,180 | +40% | 0 | 0 | — |
case-18 | fail→pass | 4,054 | 3,411 | -16% | 1 | 1 | 0% | 766 | 2,413 | +215% | 0 | 0 | — |
case-19 | pass→pass | 7,168 | 5,438 | -24% | 1 | 1 | 0% | 1,421 | 2,841 | +100% | 0 | 0 | — |
case-20 | pass→pass | 8,219 | 4,771 | -42% | 1 | 1 | 0% | 1,548 | 2,771 | +79% | 0 | 0 | — |
case-21 | pass→pass | 2,769 | 2,170 | -22% | 1 | 1 | 0% | 471 | 2,173 | +361% | 0 | 0 | — |
case-22 | pass→pass | 2,557 | 2,674 | +5% | 1 | 1 | 0% | 495 | 2,379 | +381% | 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 +36 percentage points is the difference between those two pass rates over the 21 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.