| name | qdash-issue-triage |
| description | Triage QDash issues and support knowledge for users through qdash-client profiles. Use when a user asks about open issues, known calibration problems, issue knowledge, task-result issues, AI review notes, forum posts, incident summaries, root-cause clues, or recommended follow-up in QDash. |
QDash Issue Triage
Use this skill to turn QDash issue and knowledge data into an actionable triage summary. Load qdash-core access rules first when credentials, profiles, or helper command syntax are not already clear.
Workflow
- Identify the scope: chip, qid/coupling, task name, date range, open/closed status, severity, or user-reported symptom.
- Start with
issues, issue-knowledge, task-result-issues, and ai-reviews.
- Pull supporting evidence from
task-result, task-note, task-knowledge, qubit-latest, coupling-latest, and recent history when needed.
- Group related findings by symptom, target, task, or likely cause.
- Recommend concrete next steps: inspect records, compare history, run a workflow, create/update an issue, or approve/reject knowledge. Ask before any write action.
Helper Commands
uv run --with qdash-client python skills/qdash/qdash-core/scripts/qdash_query.py --profile local issues --limit 20 --is-closed false
uv run --with qdash-client python skills/qdash/qdash-core/scripts/qdash_query.py --profile local issue-knowledge --status approved --limit 20
uv run --with qdash-client python skills/qdash/qdash-core/scripts/qdash_query.py --profile local task-result-issues --task-id task-001
uv run --with qdash-client python skills/qdash/qdash-core/scripts/qdash_query.py --profile local ai-reviews --chip-id chip-001 --limit 20
uv run --with qdash-client python skills/qdash/qdash-core/scripts/qdash_query.py --profile local task-note --task-id task-001
Output
- Start with priority: urgent, needs follow-up, informational, or resolved.
- Include issue IDs, task IDs, review IDs, affected targets, and dates.
- Call out uncertainty when evidence is circumstantial.
- Do not close, reopen, extract knowledge, approve, reject, or post replies without explicit user confirmation.