FullPipeline Workflow
End-to-end backlog triage: fetch → analyze → report.
Steps
1. Fetch (if needed)
Check if srvkp-backlog-full.json exists and is less than 24 hours old. If stale or missing:
- Run
workflows/Fetch.md
2. Analyze
Run workflows/Analyze.md with resume support:
- Load existing analysis checkpoint
- Only process unanalyzed issues
- Save checkpoints every ~20 issues
3. Report
Run workflows/Report.md:
- Generate HTML report
- Open in browser
Options
The user may specify:
| Option | Effect |
|---|---|
--component "Pipelines as Code" |
Only analyze issues for that component |
--shallow |
Quick heuristic classification, no upstream investigation |
--deep |
Full investigation for every issue (slow but thorough) |
--close-only |
Only analyze issues likely to be closable (old, EOL, no-description) |
--reset |
Discard previous analysis, re-analyze everything |
--parallel |
Use subagents for parallel processing |
Typical Run Times
| Scope | Depth | Issues | Estimated Time |
|---|---|---|---|
| Full backlog | shallow | ~1500 | 5-10 minutes |
| Full backlog | medium | ~1500 | 30-60 minutes |
| Close candidates | deep | ~300 | 15-30 minutes |
| Single component | deep | ~100 | 10-15 minutes |