| name | slope-retro |
| description | Run SLOPE post-merge retros from Codex and persist durable learnings. Trigger after a pull request merges, when the user asks for a retro, or when sprint closeout should save lessons into memory. |
SLOPE Post-Merge Retro
Use this after the PR is merged, or when the user asks to preserve lessons from a completed sprint or PR.
Prepare
- Confirm the sprint number from
slope sprint status, the roadmap, the scorecard, or the user.
- Capture the merged PR number when available.
- Identify a short summary, durable learnings, hazards, and follow-ups. Do not include secrets or private tokens.
Run
- Preview when details are uncertain:
slope retro post-merge --sprint=<N> --pr=<PR> --summary="<summary>" --learning="<learning>" --dry-run
- Persist the retro:
slope retro post-merge --sprint=<N> --pr=<PR> --summary="<summary>" --learning="<learning>" --hazard="<hazard>" --follow-up="<follow-up>"
- Repeat
--learning, --hazard, and --follow-up as needed.
- Use
--learning=project:8:<text> or another supported category[:weight]:text prefix when the memory category matters. Supported categories are workflow, style, project, hazard, and other; process is accepted as an alias for workflow; weight must be 1-10.
Verify
- Check command output for the saved
.slope/retros/post-merge/sprint-N[-pr-M].json path.
- Confirm memory persistence with
slope memory search "<distinct learning text>".
- If the command reports skipped memories, treat them as duplicate-safe persistence unless the user expected a new learning.
Report
Tell the user the retro path, memory rows added/skipped, and any follow-up work captured. If a follow-up should become roadmap work, add or triage it before ending the phase.
