Mission

Objective

Define bounded canonical metadata adoption, obtain a real supersession-edge sample, and validate reciprocal and semantic provenance checks.

Why this matters

RFR-009 tooling exists, but low metadata coverage and no real supersession edge prevent reliable repository-wide provenance traversal.

Scope

Included: adoption policy, a bounded current-artifact migration, real edge sample, heading/bare-path/registry-reference decisions, seeded tests, and retrieval impact measurement. Excluded: indiscriminate legacy migration.

Existing context

research/frontier/records/RFR-009.md and both RFR-009 investigation records.

Initial hypotheses

Targeted metadata on current canonical artifacts will improve traversal enough to justify its maintenance cost. The competing result is negligible retrieval benefit or excessive migration burden.

Required evidence

Before/after inventory, policy decision, migrated sample, real supersession records, seeded defects, validator results, and retrieval/handoff observations.

Constraints

Preserve legacy IDs and provenance; do not enforce fields before policy approval; do not invent supersession relationships.

Execution instructions

Select the smallest decision-relevant artifact class, measure baseline, migrate, validate, and compare retrieval/handoff behavior.

Deliverables

Adoption decision, migration map if needed, calibrated checks, updated inventory, and RFR-009 disposition.

Success criteria

Zero unexplained canonical reference failures, all seeded defects detected, and a defensible adopt/limit/reject decision for metadata expansion.

Stop conditions

Stop if no real provenance edge exists or measured value cannot justify broader migration.

Handoff requirements

State whether RFR-009 can close and which artifact classes remain diagnostic.