[Pepps233]

Summary

Measure procedural-information loss at the Skill Curator and synthesized-draft boundaries while preserving persistent candidates and reviewable drafts without adding a skill writer or activation path.

Reasoning

Final lessons can retain a useful conclusion while omitting material action order, decision branches, tool or environment dependencies, failure signals, recovery actions, or verification criteria. Before expanding the target-neutral Reflector contract, the pipeline needs durable evidence showing where those omissions materially weaken candidate generation or synthesis.

Skill Curator findings must cite inspected frozen evidence and retain immutable lesson and evidence provenance. Synthesized drafts are checked deterministically against the candidate procedural contract, so incomplete or reordered procedures remain reviewable but are rejected from later lifecycle stages.

Changes

• Add immutable procedural-information-loss findings for Skill Curator results.
• Require each Curator finding to cite inspected evidence already closed by its source lessons.
• Persist loss findings in the skill-curation checkpoint alongside candidates and deferred findings.
• Require synthesized drafts to preserve candidate triggers, preconditions, ordered actions, branches, tools, environment assumptions, failure handling, recovery, and verification criteria.
• Reject incomplete or reordered drafts through deterministic validation while retaining their draft and report for review.
• Version the updated skill-curation and synthesis prompts.

User impact

Operators gain durable evidence of procedural information lost between reflection, curation, and synthesis. Candidates and drafts remain non-executable; this pull request does not install, write, or activate generated skills.

Stack

Position: 7 of 12
Base: #140
Next: #148

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
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 (this PR)
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

• npm run check
• 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 test/skill-draft-validator.test.ts test/skill-synthesizer.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.