ln-316-review-refinement-worker（ln 316 审查精修工）

Paths: File paths (shared/, references/, ../ln-*) are relative to skills repo root.

Type: L3 Worker Category: 3XX Planning

Review Refinement Worker

Mandatory Read

MANDATORY READ: Load shared/references/evaluation_worker_runtime_contract.md, shared/references/evaluation_summary_contract.md, shared/references/refinement_trace_contract.md, shared/references/cleanup_evidence_contract.md MANDATORY READ: Load shared/agents/prompt_templates/iterative_refinement.md, shared/agents/prompt_templates/refinement_perspectives.md MANDATORY READ: Load shared/references/monitor_integration_pattern.md, shared/references/agent_review_workflow.md (Step: Iterative Refinement)

Purpose

• run 2-stage refinement after merge using the non-host advisor agent via agent_runner.mjs
• Stage 1: 3 parallel independent advisor sessions (dry_run_executor, new_dev_tester, adversarial_reviewer)
• Stage 2: 1 sequential advisor session (final_sweep) after merging Stage 1 results
• record refinement trace and cleanup evidence for every advisor session

Critical: refinement launches the advisor externally. Do NOT use host-native sub-agents for this phase.

Runtime

Runtime family:

• evaluation-worker-runtime

Required manifest fields:

• identifier
• phase_order
• summary_kind=review-refinement
• operation=refinement

Recommended phase_order:

1. PHASE_0_CONFIG
2. PHASE_1_STAGE1_PARALLEL
3. PHASE_2_STAGE2_FINAL_SWEEP
4. PHASE_3_WRITE_SUMMARY
5. PHASE_4_SELF_CHECK

Refinement State Machine

Critical: Independent Sessions

Each perspective MUST be a separate node agent_runner.mjs --agent {advisor_agent} invocation. Do NOT combine multiple perspectives into a single advisor prompt or session. Each iter{N}/ subdirectory = independent advisor process with its own PID.

Perspective Classification

All 4 perspectives are MANDATORY. generic_quality is not included — it is covered by the Phase 2 advisor review (review_base.md + mode template).

Stage 1: Parallel Specialized Reviews

1. Build artifact: Read current state of reviewed artifact (Story+Tasks / plan file / context docs).
2. For EACH of 3 perspectives, in parallel: a. Load perspective from refinement_perspectives.md matching the perspective name. b. Build prompt: fill iterative_refinement.md placeholders ({artifact_type}, {artifact_content}, {project_context}, {review_perspective}, {iteration_number}, {max_iterations}, {previous_findings_summary}). c. Save prompt to .hex-skills/agent-review/refinement/{identifier}/iter{N}/prompt.md node shared/agents/agent_runner.mjs --agent {advisor_agent} \ --prompt-file.hex-skills/agent-review/refinement/{identifier}/iter{N}/prompt.md \ --output-file.hex-skills/agent-review/refinement/{identifier}/iter{N}/result.md \ --cwd {project_dir}

- iter1/ = dry_run_executor - iter2/ = new_dev_tester - iter3/ = adversarial_reviewer d. Launch independent advisor process:

1. Wait for ALL 3 via runtime sync-agent; Claude hosts may use Monitor for observability (see Waiting section below).
2. Parse results from each completed session: extract JSON from ## Structured Data section.
3. Merge findings: deduplicate by (area, issue), keep higher confidence.
4. Classify: HIGH (impact_percent >= 20%), MEDIUM (10-19%), LOW (< 10%).
5. Architecture Gate on each accepted fix: "Does this implement the correct architecture directly, without backward compatibility shims?"
6. Apply accepted fixes.
7. Kill all 3 processes: node agent_runner.mjs --verify-dead {pid} per session. MANDATORY on Windows.
8. Record cleanup evidence per cleanup_evidence_contract.md for each session.
9. Build {previous_findings_summary} for Stage 2.

If ALL 3 advisor sessions fail → EXIT(ERROR), skip Stage 2. If some fail → continue with available results, record partial errors.

Stage 2: Final Sweep

1. Build artifact: Read post-fix state after Stage 1.
2. Load final_sweep perspective from refinement_perspectives.md.
3. Build prompt with {previous_findings_summary} from Stage 1.
4. Save prompt to .hex-skills/agent-review/refinement/{identifier}/iter4/prompt.md.
5. Launch advisor (single independent session).
6. Wait via runtime sync-agent; Claude hosts may use Monitor for observability.
7. Parse result, apply any accepted fixes (Architecture Gate on each).
8. Kill process, record cleanup evidence.

Waiting for Advisor Results (MANDATORY)

Use the active runtime sync-agent command before parsing or merge gates. Monitor is optional Claude Code observability only.

For EACH launched advisor process:

When running under Claude Code, optional observability:

Monitor(command="tail -f {agent_log} | grep --line-buffered -E 'Phase|ERROR|DONE'", timeout_ms=120000, description="advisor refinement {perspective_name}")

After each sync/monitor cycle:

• Check result file for <!-- END_AGENT_REVIEW_RESULT --> marker.
• Marker present → parse result, proceed.
• Marker absent, log growing → continue runtime sync or optional monitor cycle.
• Marker absent, log stale >3 min → run Liveness Protocol (see agent_review_workflow.md).

Do NOT use sleep or manual stat-polling as the primary wait mechanism.

Process Cleanup

After each advisor call (both stages):

1. Extract pid from runner stdout or metadata.
2. Run node shared/agents/agent_runner.mjs --verify-dead {pid}.
3. Record cleanup evidence per cleanup_evidence_contract.md.
4. CLI advisor processes can accumulate on Windows if not killed.

Exit States

Summary

Emit summary_kind=review-refinement.

Payload must include:

• worker=ln-316
• status
• operation=refinement
• warnings

Prefer these fields:

• stages_completed (int: 1 or 2)
• exit_reason (enum: COMPLETED, PARTIAL_ERROR, ERROR, SKIPPED)
• applied (int: total suggestions applied across all stages)
• architecture_gate_rejections (count)
• stage1_perspectives (list of completed perspective names)
• stage1_failed (list of failed perspective names)
• metadata.refinement_trace

Definition of Done

• Stage 1: all 3 advisor sessions launched in parallel
• Stage 2: final_sweep launched after Stage 1 merge
• All advisors launched via agent_runner.mjs (not host-native sub-agents)
• Runtime sync-agent used for waiting; Claude Monitor is optional observability
• Refinement trace recorded per refinement_trace_contract.md
• Cleanup evidence recorded for all launched processes
• review-refinement summary written
• Self-check passed

Version: 2.0.0 Last Updated: 2026-04-13