Skip to content

feat(agent): curate and checkpoint skill candidates - #137

Merged
Pepps233 merged 2 commits into
feat/evolved-skill-reflection-evidencefrom
feat/evolved-skill-curation
Aug 31, 2026
Merged

Pepps233 merged 2 commits into
feat/evolved-skill-reflection-evidencefrom
feat/evolved-skill-curation

Conversation

@Pepps233

@Pepps233 Pepps233 commented Aug 27, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Generate immutable, evidence-grounded skill candidates after memory proposal validation while keeping candidate creation separate from durable registry processing and executable skill activation.

Reasoning

Reflection should not decide whether a lesson becomes a skill, but a dedicated curator can evaluate skillability once evidence and selected memory proposals are available. The curator needs bounded access to detailed evidence, adjacent traces, and existing skills so its decisions remain auditable and tied to the frozen workspace.

Candidate generation is intentionally separated from consolidation, scoring, synthesis, evaluation, and activation. This stage emits non-executable artifacts and deferred findings, allowing later lifecycle stages to make those decisions independently.

Changes

  • Add structured skill-candidate and deferred-finding contracts.
  • Add a target-neutral skill curator with bounded, audited evidence, adjacent-trace, and existing-skill inspection tools.
  • Preserve frozen workspace, lesson, evidence, trace, and memory-proposal provenance in generated candidates.
  • Run skill curation after memory validation and ranking.
  • Checkpoint candidate and deferred-finding artifacts before writer processing.
  • Wire the skill-curator prompt version and model stage into the coding-agent evolution runtime.

User impact

Completed evolution jobs can emit evidence-grounded skill candidates and deferred findings without creating or activating executable skills.

Stack

Position: 3 of 12
Base: #136
Next: #138

  1. feat(coding-agent): activate versioned evolved skills #134
  2. feat(agent): preserve target-neutral reflection evidence #136
  3. feat(agent): curate and checkpoint skill candidates #137 (this PR)
  4. feat(agent): persist and process skill candidate registry #138
  5. feat(agent): synthesize outcome-aware skill drafts #139
  6. feat(agent): validate evolved skill drafts #140
  7. feat(agent): measure procedural information loss #147
  8. feat(agent): evaluate evolved skill candidates #148
  9. feat(agent): activate evolved skill candidates #149
  10. feat(agent): observe evolved skill lifecycle outcomes #150
  11. feat(agent): define evolved skill memory handling #151
  12. feat(coding-agent): add evolved skill operator controls #152

Testing

  • cd packages/evolution && node ../../node_modules/vitest/dist/cli.js --run test/skill-curator.test.ts test/control-plane.test.ts test/proposal-ranking.test.ts
  • cd packages/coding-agent && node ../../node_modules/vitest/dist/cli.js --run test/evolution-runtime.test.ts

This pull request is AI-generated by Diffex.

- Preserves frozen workspace, lesson, evidence, trace, and memory-proposal provenance in structured candidates
- Restricts detailed evidence and skill access to bounded audited inspection tools
- Run skill curation after memory validation and ranking
- Persist non-executable candidates before writer processing
- Add bounded adjacent trace inspection
@Pepps233
Pepps233 force-pushed the feat/evolved-skill-curation branch from a52c9ac to 6d6e8ef Compare August 30, 2026 21:32
@Pepps233
Pepps233 merged commit b70ff6c into main Aug 31, 2026
0 of 4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant