RF-03 — Report results have no standard handling route

State: open
Repair shape: disposition-lifecycle design
Severity: medium
Related: RF-02

Finding

A report-kind assay completes without an outcome, becomes fresh, and retains its body in a result file. No standard selector or batch handoff returns fresh but unhandled reports. The documentation correctly distinguishes freshness from handling, but the operational route for the latter is missing.

Evidence

• The review-system guide says report completion has no outcome and freshness does not mean findings were handled.
• finalize_review_job() completes report pairs and advances their baselines.
• The registered warning selector handles verdict findings, not report bodies.

Why it matters

A successful critique can consume model work and become permanently skipped without its findings reaching an operator. This is a delivery gap, not a claim that reports should acquire verdict outcomes.

Provisional repair direction

Give reports the same separate handling lifecycle proposed for RF-02 while preserving their outcome-free protocol. The queue item should carry the report body, evidence pair, and an explicit handled/deferred state.

Done when

• A completed report appears in a documented handling selector.
• Handling does not invent PASS/WARN/FAIL semantics for report-kind results.
• Supersession and retirement define what happens to an unresolved report.
• Tests cover report completion, queueing, disposition, and supersession.