[Task][RFC]: Evaluate optional independent assessment under the RFC decision gates

[huangruiteng]

Outcome / 目标

评估可选独立判断，先完成研究决策。One task tracks this RFC's delivery; keep implementation PRs and milestone evidence here instead of creating a parallel task tree.

Canonical design: RFC. Roadmap: S11 · decision required, under #4574. Design acceptance is distinct from implementation, live qualification and promotion.

Current boundary

Source audit: main at ce3862e33 (2026-09-28). This is source/PR inspection, not a new test or live-qualification claim.

The RFC accepts discussion intake only. The optional progress-review sentinel and a separate shadow implementation proposal are not evidence that Jev is qualified or that product adoption is approved.

Work remaining

• Resolve Q1–Q4 for one evidenced caller: problem, finite question, checkpoint, data, existing-model comparator, rubric, budget and stop rules; retain defer/reject as valid outcomes.
• Only after that decision, perform the authorized comparison and use Q5–Q7 to decide a bounded pilot or no adoption.

Ownership and ongoing work

Reuse #4391 and existing progress-review/Explore owners. Claimable work now is the decision brief and source-grounded analysis, not an unapproved experiment or mandatory Jev integration.

Contribution route: design/evaluation decision work is available; research execution and adoption retain the explicit RFC gates.

Acceptance

• Freeze the rubric and error costs before results; compare full-workflow utility, false positives/negatives, cost and latency with fair inputs.
• D7 ranking cannot escape the legal same-priority Todo cohort; D8 cannot override fixed research budgets or authority.
• Default-off behavior, replay and removal remain intact; advice never grants execution/settlement authority.
• Reconcile the RFC's current delivery checkpoint and this issue with the integrated revision, commands, passed/failed/untested evidence and remaining gates. Close the accepted scope only; no claim that a merged PR alone completes the RFC.

Starting points and delivery boundary

• loopx/capabilities/progress_review/README.md
• packages/loopx-jev
• loopx/capabilities/explore

Base: latest main. Reuse the existing typed owner and provider boundaries. Include affected CLI/frontend/Lark companions; verify real entrypoints and backend where changed. Preserve existing first-screen review and maintainer merge gates. Public artifacts contain only synthetic/public-safe evidence, no private operational state. This task does not authorize provider promotion, benchmark launches, release/deployment or unrelated protected effects.