learnings-summary

Internal /ace:run step — Summarise what a completed run taught; seed the next PDD.

SKILL.md

GitHub

Learnings Summary

Synthesize all information from the completed opportunity into actionable learnings.

Inputs

Pipeline-only: summarises one completed run from that run's artifacts (§ Process).

Process

  1. Read all opportunity artifacts from GDrive:

    • PDD (original plan)
    • Test results
    • Monitoring reports from ACE/<opp-name>/monitoring/
    • Data reviews from ACE/<opp-name>/data-reviews/
    • OCS transcripts (LLO questions and issues)
    • LLO feedback from ACE/<opp-name>/runs/<run-id>/10-closeout/llo-feedback.md
    • Comms log
  2. Analyze against original PDD:

    • What worked as designed?
    • What didn't work or needed adjustment?
    • Were success metrics met?
    • What was unexpected?
  3. Synthesize learnings:

    • Process learnings — what to change about the ACE process itself
    • Content learnings — what to change about the intervention design
    • Technical learnings — what to change about the apps or configuration
    • Relationship learnings — what to change about LLO engagement
  4. Determine if iteration is warranted:

    • If yes, draft a new PDD incorporating the learnings
    • This new PDD can trigger another ACE cycle
  5. Write to GDrive:

    • ACE/<opp-name>/runs/<run-id>/10-closeout/learnings-summary.md — full learnings document
    • ACE/<opp-name>/runs/<run-id>/10-closeout/learnings-summary_new-pdd.md — new PDD if iteration warranted
  6. Write phases.closeout.products.learnings to the current run's run_state.yaml so downstream readers (ace-web summary, next-cycle idea-to-pdd if iteration is warranted) can deep-link to the learnings doc and the iteration PDD without listing the closeout folder.

    phases:
      closeout:
        products:
          learnings:
            summary_file_id: <Drive fileId of learnings-summary.md>
            summary_web_view_link: <webViewLink returned by Step 5's drive_create_file for learnings-summary.md>
            new_pdd_file_id: <Drive fileId of learnings-summary_new-pdd.md, or null when iteration not warranted>
            new_pdd_web_view_link: <webViewLink for the new-pdd doc, or null>
            iteration_warranted: <true | false>
    

    Both *_web_view_link fields are populated from the webViewLink returned by Step 5's drive_create_file calls. The Drive URL shape for Google Docs (/document/d/<id>/edit) differs from the generic blob preview URL (/file/d/<id>/view), and there's no single construction rule that works for every file type — so the producer skill (which has the live webViewLink in hand) records it directly, and consumers don't have to guess.

    Apply via mcp__plugin_ace_ace-gdrive__update_yaml_file with merge: 'deep' (partial phase-child patch — deep preserves the phase's status/steps and sibling products; two-level would drop them, #572/#587). Sole writer of products.learnings.

MCP Tools Used

  • Google Drive: drive_read_file, drive_create_file, drive_list_folder
  • OCS: ocs_list_sessions, ocs_get_session

Mode Behavior

  • Auto: Generate learnings, create new PDD if warranted, notify admin group
  • Review: Present learnings and new PDD for team discussion

Dry-Run Behavior

When --dry-run is active:

  • Learnings analysis and new PDD generation proceed normally (written to GDrive)
  • Write any notification emails (to admin group) to comms-log/dry-run-learnings-summary.md instead of sending
  • State tracks as dry-run-success