Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →oracle/oci-functions-troubleshoot
Diagnose OCI Functions setup, control-plane, deployment, invocation, and observability issues using a deterministic triage workflow.
Passed 15 of 21 test cases with the skill, and 3 of 21 without it.
and 4 more in the full report
16 fixed0 passed either way3 still failing3 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 16 test cases the skill fixed. Same model, same prompt, with and without the skill.
During `fn deploy --app inventory-app`, the command fails with `func.yaml: key 'schema_version' not found`. Format triage output as: stage, evidence, most_likely_cause, confidence, next_confirming_check, smallest_safe_remediation, validation.
| Checked | Without skill | With skill |
|---|---|---|
| Specifies stage as deploy | Missed | Passed |
| Identifies func.yaml schema or key syntax issue as primary cause | Passed | Passed |
| Recommends inspecting local func.yaml | Passed | Passed |
The instructions your model reads when the skill loads.
Use this skill to isolate the most likely cause of an OCI Functions problem from error text, command output, or observability signals.
This skill is diagnosis-first and non-mutating by default:
$oci-functions-deployAlways answer in this exact shape:
stage: one of setup_or_control_plane, create, deploy, invoke, observability_firstevidence: 1-3 concrete facts from the user's error text, command, config, logs, metrics, or limitsmost_likely_cause: one primary cause, or a ranked top-2 shortlist when confidence is lowconfidence: low, medium, or highnext_confirming_check: the single best read-only check to narrow or confirm the causesmallest_safe_remediation: the least invasive next action; if evidence is weak, this can be "collect more evidence with ..."validation: the follow-up step that proves the remediation workedreferences/error-patterns.md.setup_or_control_plane: local Fn/OCI/OCIR/tooling/auth/context/policy problems before normal app or function operations succeedcreate: application creation failuresdeploy: fn deploy, image build or push, func.yaml, app config, subnet annotation, or registry-target failuresinvoke: runtime invocation, HTTP, image pull, timeout, memory, networking, detached delivery, or response-limit failures after a function has been deployedobservability_first: the user starts from logs, traces, metrics, limits, or "why is this function unhealthy?"Capture the minimum facts before diagnosing:
test cases passed, vs 3 / 21 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.