| name | sh-qa |
| description | Wrapper skill that invokes the superharness qa CLI command from within the AI tool session. |
QA Evaluation
Trigger external QA evaluation by calling the superharness qa CLI command.
This skill is a thin wrapper — it runs the CLI and reports results. The actual QA logic lives in the CLI.
Process
Step 1: Run QA
superharness qa --task .superharness/tasks/{task}
This will:
- Call configured external QA services (managed HTTP / autonomous command)
- Write
qa-report.md (human-readable) to the task directory
- Write
qa-issues.json (machine-readable) to the task directory
Step 2: Report Results
Read and summarize qa-report.md for the user:
- Total checks run
- Passed / Failed counts
- List of issues by severity
Step 3: Suggest Next Action
If issues were found:
"QA found {N} issues ({critical} critical, {major} major). Run /superharness:sh-fix to address them."
If all passed:
"QA passed. Ready to finalize with superharness:finishing-a-development-branch."
If no external QA services configured:
"No external QA services configured. Feedback layer checks (test/lint/typecheck) already ran during implementation. To add QA services, configure them in .superharness/config.yaml under qa.services."