diff --git a/conductor/tracks/meta_tooling_workflow_review_20260620/plan.md b/conductor/tracks/meta_tooling_workflow_review_20260620/plan.md
new file mode 100644
index 00000000..cecc8245
--- /dev/null
+++ b/conductor/tracks/meta_tooling_workflow_review_20260620/plan.md
@@ -0,0 +1,1593 @@
+# Meta-Tooling Workflow Review — Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Produce a ≥4,000 LOC research-only analysis of past-month LLM agent behavior in Manual Slop, identifying workflow improvements the user can apply in a near-future conservative rebuild track.
+
+**Architecture:** Tier 1 Orchestrator owns the synthesis; 3 parallel Tier 3 sub-agents sweep the corpus (reports / structured data / hidden notes). Output is 7 reference artifacts at `conductor/tracks/meta_tooling_workflow_review_20260620/` — 4-part report + 5 side artifacts + 2 standalone inputs for the rebuild track. No `src/` / `tests/` / `conductor/` / `AGENTS.md` / `.opencode/` / `scripts/audit_*.py` changes.
+
+**Tech Stack:** Markdown research artifacts, git + git notes, Tier 1 inline execution with Tier 3 sub-agent dispatch via `uv run python scripts/mma_exec.py --role tier3-worker`.
+
+**Reference spec:** `conductor/tracks/meta_tooling_workflow_review_20260620/spec.md`
+
+---
+
+## File Structure
+
+### New files (12)
+
+| File | Responsibility |
+|---|---|
+| `conductor/tracks/meta_tooling_workflow_review_20260620/spec.md` | Design intent (already written) |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/metadata.json` | Structured track metadata |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/state.toml` | Track state (`current_phase` + task tracking) |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/report.md` | 4-part synthesis report (≥4,000 LOC) |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md` | ~50-row flat track reference |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md` | Prioritized rebuild backlog (15-25 entries) |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md` | Tier 3 sweep A output (per-track summaries) |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md` | Tier 3 sweep B + C output (12-pattern evidence) |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md` | Bridge to 4 sibling reviews |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md` | Standalone Part 3 input |
+| `conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md` | Standalone Part 4 input |
+
+### Modified files (1)
+
+| File | Modification |
+|---|---|
+| `conductor/tracks.md` | Add track to "Active Tracks" table in Phase 1; mark "Recently Completed" in Phase 11 |
+
+### Commit count budget
+
+13-15 atomic commits across 11 phases (1 setup + 3 sweep outputs + 1 Part 1 + 1-2 Part 2 + 1 Part 3 + 1 Part 4 + 5 side artifacts + 0-1 self-review fix + 1 finalize).
+
+---
+
+## Phase 1: Setup (1 commit)
+
+**Scope:** Create track directory; write all skeleton files; register track in `conductor/tracks.md`.
+
+### Task 1.1: Create track artifacts (1 commit)
+
+**Files:**
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/spec.md` (already exists)
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/metadata.json`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/state.toml`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/report.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md`
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md`
+- Modify: `conductor/tracks.md` (add 1 row to Active Tracks table)
+
+- [ ] **Step 1: Verify track directory exists**
+
+Run: `Test-Path -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620"`
+Expected: `True`
+
+If `False`, run:
+```powershell
+New-Item -ItemType Directory -Path "conductor\tracks\meta_tooling_workflow_review_20260620" -Force
+```
+
+- [ ] **Step 2: Write `metadata.json`**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/metadata.json`:
+
+```json
+{
+  "track_id": "meta_tooling_workflow_review_20260620",
+  "name": "Meta-Tooling Workflow Review — Past-Month LLM Behavior Analysis",
+  "type": "research-only",
+  "priority": "medium-high",
+  "owner": "Tier 1 Orchestrator (sole synthesis author); Tier 3 sub-agents for parallel sweeps",
+  "initialized": "2026-06-20",
+  "status": "active",
+  "blocked_by": [],
+  "blocks": ["workflow_improvements_rebuild_<future-date>"],
+  "scope": {
+    "new_files": [
+      "conductor/tracks/meta_tooling_workflow_review_20260620/spec.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/metadata.json",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/state.toml",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/report.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md",
+      "conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md"
+    ],
+    "modified_files": [
+      "conductor/tracks.md"
+    ],
+    "deleted_files": []
+  },
+  "user_directives": [
+    {"date": "2026-06-20", "directive": "Full past month (~75 reports + git log + state.toml + guide docs)", "source": "user (this session, Q1 confirmation)"},
+    {"date": "2026-06-20", "directive": "Document-driven (4 parts): What shipped / LLM Behavior Patterns / Workflow Improvements / Implementation Sequencing", "source": "user (Q2 confirmation)"},
+    {"date": "2026-06-20", "directive": "Audit depth C: reports + git log + track spec deviations + state.toml + guide docs", "source": "user (Q3 confirmation)"},
+    {"date": "2026-06-20", "directive": "Recommendation structure D: by target doc × by confidence tier", "source": "user (Q4 confirmation)"},
+    {"date": "2026-06-20", "directive": "Execution model C: Tier 1 anchor + Tier 3 parallel sweeps; sub-agents for batch data only", "source": "user (Q5 confirmation)"},
+    {"date": "2026-06-20", "directive": "Output shape C: report + side artifacts + workflow_improvements.md + implementation_sequencing.md", "source": "user (Q6 confirmation)"},
+    {"date": "2026-06-20", "directive": "Minimum 4,000 line report; use nagent_review_v3.1 chunking strategy", "source": "user (Q7 confirmation)"},
+    {"date": "2026-06-20", "directive": "Be conservative with meta-tooling to not break OpenCode", "source": "user (overall framing)"}
+  ],
+  "verification_criteria": [
+    "report.md has all 4 parts present and non-empty",
+    "report.md total LOC >= 4,000",
+    "Part 1 has all 5 track-family sub-sections",
+    "Part 2 has 8-16 LLM behavior patterns (target 12) with the 7-sub-section structure + verdict block",
+    "Part 3 has 15-25 workflow improvements organized by 5 target docs x 3 confidence tiers",
+    "Part 4 has all 5 implementation phases with the 5-sub-section structure",
+    "comparison_table.md has ~50 rows",
+    "decisions.md has 15-25 entries sorted HIGH to LOW with destination files",
+    "shipped_work_index.md exists with per-track summaries",
+    "llm_behavior_catalog.md exists with the 12-pattern catalog",
+    "nagent_takeaways_meta_tooling_20260620.md exists with 5-part bridge structure",
+    "workflow_improvements.md exists as standalone",
+    "implementation_sequencing.md exists as standalone",
+    "No src/ / tests/ / AGENTS.md / conductor/*.md / .opencode/ / scripts/audit_*.py changes",
+    "User has reviewed and approved the final report",
+    "conductor/tracks.md updated",
+    "All atomic commits have git notes attached",
+    "state.toml final state is current_phase=11 and status=active"
+  ],
+  "regressions_and_pre_existing_failures": [],
+  "pre_existing_failures_remaining": [],
+  "deferred_to_followup_tracks": [
+    {
+      "title": "Workflow Improvements Rebuild",
+      "description": "Apply the 5-phase conservative sequencing from Part 4 to AGENTS.md / conductor/workflow.md / conductor/code_styleguides/error_handling.md / .opencode/agents/*.md / scripts/audit_*.py",
+      "track_status": "planned"
+    }
+  ]
+}
+```
+
+- [ ] **Step 3: Write `state.toml`**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/state.toml`:
+
+```toml
+# Track state for meta_tooling_workflow_review_20260620
+# Updated by Tier 1 Orchestrator as tasks complete
+
+[meta]
+track_id = "meta_tooling_workflow_review_20260620"
+name = "Meta-Tooling Workflow Review — Past-Month LLM Behavior Analysis"
+status = "active"
+current_phase = 1
+last_updated = "2026-06-20"
+
+[blocked_by]
+# No blockers
+
+[blocks]
+# Future workflow-improvements rebuild track
+workflow_improvements_rebuild = "planned in meta_tooling_workflow_review_20260620"
+
+[phases]
+phase_1 = { status = "in_progress", checkpointsha = "", name = "Setup" }
+phase_2 = { status = "pending", checkpointsha = "", name = "Tier 3 sub-agent sweeps" }
+phase_3 = { status = "pending", checkpointsha = "", name = "Tier 1 anchor read" }
+phase_4 = { status = "pending", checkpointsha = "", name = "Part 1 — What Shipped" }
+phase_5 = { status = "pending", checkpointsha = "", name = "Part 2 — LLM Behavior Patterns" }
+phase_6 = { status = "pending", checkpointsha = "", name = "Part 3 — Workflow Improvements" }
+phase_7 = { status = "pending", checkpointsha = "", name = "Part 4 — Implementation Sequencing" }
+phase_8 = { status = "pending", checkpointsha = "", name = "Side artifacts + standalone inputs" }
+phase_9 = { status = "pending", checkpointsha = "", name = "Self-review" }
+phase_10 = { status = "pending", checkpointsha = "", name = "User review gate" }
+phase_11 = { status = "pending", checkpointsha = "", name = "Finalize" }
+
+[tasks]
+# Phase 1 — Setup
+t1_1_setup_artifacts = { status = "in_progress", commit_sha = "", description = "Create all skeleton files + register in tracks.md" }
+
+# Phase 2 — Tier 3 sub-agent sweeps (3 parallel)
+t2_1_sweep_a_reports = { status = "pending", commit_sha = "", description = "Tier 3 sweep A: reports corpus → shipped_work_index.md" }
+t2_2_sweep_b_structured = { status = "pending", commit_sha = "", description = "Tier 3 sweep B: git log + state.toml + spec deviations → llm_behavior_catalog.md (part 1)" }
+t2_3_sweep_c_hidden_notes = { status = "pending", commit_sha = "", description = "Tier 3 sweep C: guide docs + AGENTS.md + conductor/*.md → llm_behavior_catalog.md (appended)" }
+
+# Phase 3 — Tier 1 anchor read (no commits; internal scratchpad)
+t3_1_anchor_read = { status = "pending", commit_sha = "", description = "Read 10 anchor reports; produce internal scratchpad" }
+
+# Phase 4 — Part 1 synthesis
+t4_1_part1_synthesis = { status = "pending", commit_sha = "", description = "Write Part 1 (5 sub-sections × 160-200 LOC each)" }
+
+# Phase 5 — Part 2 synthesis
+t5_1_part2_synthesis = { status = "pending", commit_sha = "", description = "Write Part 2 (12 patterns × 125-170 LOC each, 7-sub-section structure)" }
+
+# Phase 6 — Part 3 synthesis
+t6_1_part3_synthesis = { status = "pending", commit_sha = "", description = "Write Part 3 (15-25 improvements × 50-80 LOC each, by target doc × confidence tier)" }
+
+# Phase 7 — Part 4 synthesis
+t7_1_part4_synthesis = { status = "pending", commit_sha = "", description = "Write Part 4 (5 phases × 60-100 LOC each, conservative sequencing)" }
+
+# Phase 8 — Side artifacts + standalone inputs (5 commits)
+t8_1_comparison_table = { status = "pending", commit_sha = "", description = "Write comparison_table.md (~50 rows)" }
+t8_2_decisions = { status = "pending", commit_sha = "", description = "Write decisions.md (15-25 entries)" }
+t8_3_nagent_takeaways = { status = "pending", commit_sha = "", description = "Write nagent_takeaways_meta_tooling_20260620.md (5-part bridge)" }
+t8_4_workflow_improvements_standalone = { status = "pending", commit_sha = "", description = "Write workflow_improvements.md (Part 3 verbatim standalone)" }
+t8_5_implementation_sequencing_standalone = { status = "pending", commit_sha = "", description = "Write implementation_sequencing.md (Part 4 verbatim + phase dependencies)" }
+
+# Phase 9 — Self-review (0-1 commits)
+t9_1_self_review = { status = "pending", commit_sha = "", description = "Placeholder scan + internal consistency + scope check + ambiguity check + chunking verification; fix inline" }
+
+# Phase 10 — User review gate (0 commits; user-driven)
+t10_1_user_review = { status = "pending", commit_sha = "", description = "User reviews report + side artifacts + standalone inputs; approves or iterates" }
+
+# Phase 11 — Finalize
+t11_1_finalize = { status = "pending", commit_sha = "", description = "Update state.toml to current_phase=11; update metadata.json with final stats; mark Recently Completed in tracks.md" }
+
+[verification]
+phase_1_complete = false
+phase_2_complete = false
+phase_3_complete = false
+phase_4_complete = false
+phase_5_complete = false
+phase_6_complete = false
+phase_7_complete = false
+phase_8_complete = false
+phase_9_complete = false
+phase_10_complete = false
+phase_11_complete = false
+report_4k_loc_floor_met = false
+user_review_approved = false
+```
+
+- [ ] **Step 4: Write `report.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/report.md`:
+
+```markdown
+# Meta-Tooling Workflow Review — Past-Month LLM Behavior Analysis
+
+**Track:** `meta_tooling_workflow_review_20260620`
+**Shipped:** (pending)
+**Owner:** Tier 1 Orchestrator (sole synthesis author)
+**Date:** 2026-06-20
+**Scope:** Past month (2026-05-20 → 2026-06-20)
+**Status:** Research-only. No production code changes.
+
+---
+
+## Part 1 — What Shipped
+
+### §1.1 The Result Migration Campaign (5 sub-tracks + umbrella)
+
+[160-200 LOC; populated in Phase 4]
+
+### §1.2 Tier 2 Autonomous Sandbox Family
+
+[160-200 LOC; populated in Phase 4]
+
+### §1.3 Stability & Test-Infrastructure Tracks
+
+[160-200 LOC; populated in Phase 4]
+
+### §1.4 Meta-Analysis Corpus
+
+[160-200 LOC; populated in Phase 4]
+
+### §1.5 One-Off Fixes & Polishes
+
+[160-200 LOC; populated in Phase 4]
+
+---
+
+## Part 2 — LLM Behavior Patterns
+
+### §2.1 Anti-Sliming (Heuristic Laundering)
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.2 Hard-Gate Bypass (Manual Review → Bulk Verify)
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.3 Regression-After-Refactor (Lost Context in Extraction)
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.4 Heuristic Proliferation Mid-Track
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.5 Tier 2 Escalation Drift (Ambiguous User Intent)
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.6 Report-As-Substitute-For-Fix
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.7 Decision-Deflection ("Not Going To Attempt Another Fix")
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.8 Lost-Context Extraction
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.9 Literal-vs-Inferred Instruction Interpretation
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.10 Cross-Track Synthesis Gap
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.11 The "I'm Done" Surrender Threshold
+
+[125-170 LOC; populated in Phase 5]
+
+### §2.12 Anti-Sliming Protocol Evolution (Phase 10→11→12→13)
+
+[125-170 LOC; populated in Phase 5]
+
+---
+
+## Part 3 — Workflow Improvements
+
+### §AGENTS.md — Anti-Patterns + Hard Rules
+
+[populated in Phase 6]
+
+### §conductor/workflow.md — Process Gates
+
+[populated in Phase 6]
+
+### §conductor/code_styleguides/error_handling.md — Convention Tightening
+
+[populated in Phase 6]
+
+### §.opencode/agents/tier2-autonomous.md — Tier-Specific Directives
+
+[populated in Phase 6]
+
+### §scripts/audit_*.py — Enforcement Script Updates
+
+[populated in Phase 6]
+
+---
+
+## Part 4 — Implementation Sequencing
+
+### §Phase 1: AGENTS.md Doc Edits (Zero Risk)
+
+[populated in Phase 7]
+
+### §Phase 2: conductor/workflow.md Process Gate Additions
+
+[populated in Phase 7]
+
+### §Phase 3: conductor/code_styleguides/error_handling.md Convention Tightening
+
+[populated in Phase 7]
+
+### §Phase 4: .opencode/ Tier-Specific Directive Updates
+
+[populated in Phase 7]
+
+### §Phase 5: scripts/audit_*.py + CI Gate Additions
+
+[populated in Phase 7]
+```
+
+- [ ] **Step 5: Write `comparison_table.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md`:
+
+```markdown
+# Comparison Table — Past-Month Tracks
+
+**Date:** 2026-06-20
+**Scope:** ~75 tracks active or shipped in the past month (2026-05-20 → 2026-06-20)
+
+| Track family | Track name | Status | Key reports | First LLM-behavior tag |
+|---|---|---|---|---|
+```
+
+- [ ] **Step 6: Write `decisions.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md`:
+
+```markdown
+# Decisions — Workflow Improvements Backlog
+
+**Date:** 2026-06-20
+**Scope:** Conservative workflow improvements for the near-future rebuild track
+**Priority order:** HIGH → MEDIUM → LOW
+
+| # | Priority | Workflow improvement | Change | Destination file | Evidence | Risk | Sequencing phase |
+|---|---|---|---|---|---|---|---|
+```
+
+- [ ] **Step 7: Write `shipped_work_index.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md`:
+
+```markdown
+# Shipped Work Index — Past-Month Tracks
+
+**Date:** 2026-06-20
+**Generated by:** Tier 3 sub-agent sweep A
+**Source:** `docs/reports/*.md` past month + `conductor/tracks.md` Active Tracks table + `git log --oneline`
+
+---
+
+[populated by Tier 3 sweep A; per-track summaries ~10-20 LOC each × ~50 tracks = 300-500 LOC]
+```
+
+- [ ] **Step 8: Write `llm_behavior_catalog.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md`:
+
+```markdown
+# LLM Behavior Catalog — Past-Month Evidence
+
+**Date:** 2026-06-20
+**Generated by:** Tier 3 sub-agent sweeps B + C
+**Sources:** git log + git notes + state.toml user_directives + spec.md deviation sections + docs/guide_*.md
+
+---
+
+[populated by Tier 3 sweep B; 12 patterns × 30-50 LOC each = 360-600 LOC]
+[appended by Tier 3 sweep C; hidden-notes section ~200 LOC]
+```
+
+- [ ] **Step 9: Write `nagent_takeaways_meta_tooling_20260620.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md`:
+
+```markdown
+# Nagent Takeaways — Meta-Tooling Workflow Review Bridge
+
+**Date:** 2026-06-20
+**Bridges:** `meta_tooling_workflow_review_20260620` ↔ `nagent_review_20260608` ↔ `fable_review_20260617` ↔ `superpowers_review_20260619` ↔ `intent_dsl_survey_20260612`
+
+---
+
+## TL;DR
+
+[1 paragraph; populated in Phase 8]
+
+## Cross-Reference Table
+
+[~10-15 rows; one row per LLM pattern that touches a verdict in the sibling reviews]
+
+## The N New Findings This Track Adds
+
+[populated in Phase 8]
+
+## The M Sibling-Review Findings This Track Contradicts or Extends
+
+[populated in Phase 8]
+
+## Pointer to fable_review
+
+[1 paragraph; populated in Phase 8]
+```
+
+- [ ] **Step 10: Write `workflow_improvements.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md`:
+
+```markdown
+# Workflow Improvements — Standalone Catalog
+
+**Date:** 2026-06-20
+**Purpose:** Standalone input for the near-future "workflow improvements rebuild" track. Read this WITHOUT re-reading the 4,000-LOC `report.md`.
+
+---
+
+[verbatim copy of `report.md` Part 3, populated in Phase 8]
+```
+
+- [ ] **Step 11: Write `implementation_sequencing.md` skeleton**
+
+Write to `conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md`:
+
+```markdown
+# Implementation Sequencing — Standalone Plan
+
+**Date:** 2026-06-20
+**Purpose:** Standalone input for the near-future "workflow improvements rebuild" track. Read this WITHOUT re-reading the 4,000-LOC `report.md`.
+
+---
+
+[verbatim copy of `report.md` Part 4 + phase dependencies section, populated in Phase 8]
+```
+
+- [ ] **Step 12: Update `conductor/tracks.md`**
+
+Open `conductor/tracks.md` and add a row to the "Active Tracks" table (the table at the top of the file). Insert the following row after the existing rows (find the last A-priority row):
+
+```
+| 22b | A (meta-tooling) | [Meta-Tooling Workflow Review — Past-Month LLM Behavior Analysis](#track-meta-tooling-workflow-review-past-month-llm-behavior-analysis) | spec ✓, plan ✓, **active** | (none — independent; **NEW 2026-06-20**; sibling to nagent_review + fable_review + superpowers_review + intent_dsl_survey; produces workflow_improvements.md + implementation_sequencing.md as standalone inputs for a near-future rebuild track; research-only; no src/, tests/, AGENTS.md, conductor/*.md, .opencode/, or scripts/audit_*.py changes) |
+```
+
+Also append a section at the bottom of "Active Tracks":
+
+```
+### Track: Meta-Tooling Workflow Review — Past-Month LLM Behavior Analysis
+
+*Spec: [./tracks/meta_tooling_workflow_review_20260620/spec.md](./tracks/meta_tooling_workflow_review_20260620/spec.md), Plan: [./tracks/meta_tooling_workflow_review_20260620/plan.md](./tracks/meta_tooling_workflow_review_20260620/plan.md)*
+*Goal: Analyze all past-month reports + git log + state.toml + guide docs; identify 8-16 LLM behavior patterns; catalog workflow improvements by target doc × confidence tier; sequence for a conservative rebuild. 4-part report ≥4,000 LOC. Research-only; recommendations go to workflow_improvements.md + implementation_sequencing.md for the future rebuild track.*
+*Status: in progress (Phase 1 setup complete; Tier 3 sweeps pending).*
+```
+
+- [ ] **Step 13: Verify all 11 files exist**
+
+Run: `Get-ChildItem -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620" -File | Select-Object Name, Length | Format-Table -AutoSize`
+Expected: 11 files (spec.md + metadata.json + state.toml + report.md + comparison_table.md + decisions.md + shipped_work_index.md + llm_behavior_catalog.md + nagent_takeaways_meta_tooling_20260620.md + workflow_improvements.md + implementation_sequencing.md)
+
+- [ ] **Step 14: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/ conductor/tracks.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 1 setup (11 skeleton files + tracks.md registration)"
+```
+
+- [ ] **Step 15: Attach git note**
+
+```bash
+git notes add -m "Spec + metadata + state + 11 skeleton files for meta_tooling_workflow_review_20260620. Research-only track analyzing past-month LLM agent behavior (2026-05-20 to 2026-06-20). 4-part report (>=4,000 LOC target). Tier 1 anchor + 3 Tier 3 sub-agent sweeps. 7 reference artifacts (report + 5 side artifacts + 2 standalone inputs for future rebuild track). 11 phases. Sibling to nagent_review / fable_review / superpowers_review / intent_dsl_survey. No src/ / tests/ / conductor/*.md / AGENTS.md / .opencode/ / scripts/audit_*.py changes." $(git log -1 --format="%H")
+```
+
+- [ ] **Step 16: Update `state.toml`**
+
+Update `state.toml`:
+- `current_phase = 2`
+- `phase_1 = { status = "completed", checkpointsha = "<paste first 7 chars of commit hash>", name = "Setup" }`
+- `t1_1_setup_artifacts = { status = "completed", commit_sha = "<paste first 7 chars of commit hash>", description = "..." }`
+- Add `<paste first 7 chars of commit hash>` to the `last_updated` field
+
+Then commit:
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 1 (Setup) complete for meta_tooling_workflow_review_20260620"
+```
+
+Attach git note:
+```bash
+git notes add -m "Phase 1 complete: 11 skeleton files + tracks.md registration committed. 1 commit this phase. Phase 2 (Tier 3 sub-agent sweeps) is next." $(git log -1 --format="%H")
+```
+
+---
+
+## Phase 2: Tier 3 Sub-Agent Sweeps (3 commits)
+
+**Scope:** Dispatch 3 Tier 3 sub-agents in parallel (Sweep A: reports corpus; Sweep B: structured data; Sweep C: hidden notes). Each sub-agent returns a structured markdown output that Tier 1 reviews before committing.
+
+### Task 2.1: Sweep A — Reports corpus → `shipped_work_index.md` (1 commit)
+
+**Files:**
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md` (populated by sub-agent)
+
+- [ ] **Step 1: Verify Tier 3 dispatch mechanism**
+
+Run: `Test-Path -LiteralPath "scripts\mma_exec.py"`
+Expected: `True` (the `mma_exec.py` script must exist for Tier 3 dispatch)
+
+If `False`, STOP and report to user — Tier 3 dispatch is not available, fall back to Tier 1 sequential read of all ~75 reports (will require plan restructuring).
+
+- [ ] **Step 2: Build the Sweep A prompt**
+
+Use the following prompt (verbatim, with file paths substituted):
+
+```
+You are a Tier 3 Worker sub-agent dispatched for a meta-tooling review track.
+
+TASK: Read all ~75 markdown reports in C:/projects/manual_slop/docs/reports/ dated 2026-05-20 through 2026-06-20 (use Get-ChildItem -LiteralPath "docs/reports" -File | Where-Object { $_.LastWriteTime -ge (Get-Date).AddDays(-35) } to enumerate). For each report, extract:
+- Track folder (if the report references one)
+- Shipped date (from filename or metadata)
+- Key deliverable files referenced
+- Track family (one of: migration campaign / tier-2 sandbox / stability-test-infra / meta-analysis / one-off polish)
+- LLM-behavior touch-points (any mention of agent failure, regression, anti-pattern, deviation, or workaround)
+
+OUTPUT FORMAT: Markdown file at C:/projects/manual_slop/conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md
+
+The file structure:
+- H1 title
+- Date + scope
+- H2 section "By Track Family" with 5 sub-sections (one per family)
+- Each sub-section: bullet list of reports, each with format: `  - [report filename](path) | [track folder if any] | [shipped date] | [key deliverables] | [LLM-behavior tags]`
+- H2 section "LLM-Behavior Tag Index" (alphabetical): each tag (anti-sliming, hard-gate-bypass, etc.) → list of reports that mention it
+
+CONSTRAINTS:
+- Total output: 300-500 LOC
+- Use 1-space indentation for any nested content (per project conventions)
+- No day estimates
+- Cite specific file paths with line numbers where possible
+- If a report does not reference a track, put it in the "uncategorized" section at the bottom
+
+Return the FULL contents of shipped_work_index.md in your final message (not a summary).
+```
+
+- [ ] **Step 3: Dispatch Sweep A**
+
+Run:
+```bash
+uv run python scripts/mma_exec.py --role tier3-worker "<paste the prompt above>"
+```
+
+Expected: Sub-agent runs (may take several minutes); final message contains the full markdown content of `shipped_work_index.md`.
+
+- [ ] **Step 4: Verify output file**
+
+Run: `Get-ChildItem -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620\shipped_work_index.md" | Select-Object Length`
+Expected: file exists, length > 10,000 bytes (300-500 LOC × ~30 bytes/LOC).
+
+If length < 5,000 bytes, the sub-agent output is too thin — STOP and report to user.
+
+- [ ] **Step 5: Spot-check 5 reports**
+
+Open 5 random reports from the index. Verify:
+- Track folder is correct (where applicable)
+- LLM-behavior tags are accurate
+- File paths exist
+
+If any spot-check fails, correct the index inline before committing.
+
+- [ ] **Step 6: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Sweep A (Tier 3 reports corpus -> shipped_work_index.md)"
+```
+
+- [ ] **Step 7: Attach git note**
+
+```bash
+git notes add -m "Tier 3 Sweep A output: shipped_work_index.md (~300-500 LOC). Per-track summaries from all ~75 docs/reports/ past month. 5 track families + LLM-behavior tag index. Used as scaffolding for report.md Part 1." $(git log -1 --format="%H")
+```
+
+- [ ] **Step 8: Update `state.toml`**
+
+Update:
+- `phase_2` → 1 of 3 complete (don't bump `current_phase` yet; wait for all 3 sweeps)
+- `t2_1_sweep_a_reports = { status = "completed", commit_sha = "<hash>", description = "..." }`
+
+Commit:
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 2 Sweep A complete"
+```
+
+### Task 2.2: Sweep B — Structured data → `llm_behavior_catalog.md` (Part 1) (1 commit)
+
+**Files:**
+- Create: `conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md` (populated by sub-agent; Part 1)
+
+- [ ] **Step 1: Build the Sweep B prompt**
+
+```
+You are a Tier 3 Worker sub-agent dispatched for a meta-tooling review track.
+
+TASK: Mine 3 sources of structured data for LLM agent behavior evidence:
+
+1. GIT LOG: Run `git log --since="2026-05-20" --until="2026-06-20" --oneline` from C:/projects/manual_slop. Also run `git log --since="2026-05-20" --until="2026-06-20" --format="%H %s" --no-merges`. Identify 50-100 commits whose messages mention regression, fix, bug, anti-pattern, deviation, workaround, tier-2, autonomous, audit, heuristic, or test failure.
+
+2. GIT NOTES: Run `git log --since="2026-05-20" --until="2026-06-20" --format="%H" --no-merges | ForEach-Object { $notes = git notes show $_ 2>$null; if ($notes) { "$_`n$notes`n---" } }` from C:/projects/manual_slop. Capture all git notes attached to past-month commits.
+
+3. STATE.TOML USER_DIRECTIVES: For each conductor/tracks/<id>/state.toml in C:/projects/manual_slop/conductor/tracks/ where the track was active in the past month, read the [meta] section + user_directives_logged section. Capture all directives.
+
+4. SPEC.MD DEVIATIONS: For each conductor/tracks/<id>/spec.md in the past-month tracks, read the "Deviations from Spec/Plan" section. Capture all deviations.
+
+OUTPUT FORMAT: Markdown file at C:/projects/manual_slop/conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md
+
+File structure:
+- H1 title
+- Date + scope + sources
+- 12 sections (one per hypothesized LLM behavior pattern): Anti-Sliming / Hard-Gate Bypass / Regression-After-Refactor / Heuristic Proliferation / Tier-2 Escalation Drift / Report-As-Substitute-For-Fix / Decision-Deflection / Lost-Context Extraction / Literal-vs-Inferred Instruction / Cross-Track Synthesis Gap / "I'm Done" Surrender Threshold / Anti-Sliming Protocol Evolution
+- Each section:
+  - H3 subsection "Definition" (1-2 sentences)
+  - H3 subsection "Evidence Citations" (3-7 file:line refs with brief excerpts)
+  - H3 subsection "Status" (NEW / PARTIALLY-CODIFIED / FULLY-CODIFIED / SUBSUMED based on AGENTS.md coverage)
+  - Each subsection ~30-50 LOC
+
+CONSTRAINTS:
+- Total output: 500-700 LOC
+- Use 1-space indentation for nested content
+- No day estimates
+- Cite specific file:line for every piece of evidence
+- If a pattern has <3 evidence citations, mark it as "INSUFFICIENT EVIDENCE — consider removal in synthesis"
+
+Return the FULL contents of llm_behavior_catalog.md in your final message.
+```
+
+- [ ] **Step 2: Dispatch Sweep B**
+
+Run:
+```bash
+uv run python scripts/mma_exec.py --role tier3-worker "<paste the prompt above>"
+```
+
+- [ ] **Step 3: Verify output file**
+
+Run: `Get-ChildItem -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620\llm_behavior_catalog.md" | Select-Object Length`
+Expected: length > 15,000 bytes (500-700 LOC).
+
+If length < 10,000 bytes, the sub-agent output is too thin — STOP and report.
+
+- [ ] **Step 4: Spot-check 10 evidence citations**
+
+Open 10 random file:line citations from the catalog. Verify each exists at the cited location and the excerpt is accurate.
+
+If any spot-check fails, correct the catalog inline.
+
+- [ ] **Step 5: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Sweep B (Tier 3 structured data -> llm_behavior_catalog.md Part 1)"
+```
+
+- [ ] **Step 6: Attach git note + commit state**
+
+```bash
+git notes add -m "Tier 3 Sweep B output: llm_behavior_catalog.md Part 1 (~500-700 LOC). Evidence from git log + git notes + state.toml user_directives + spec.md deviations. 12 LLM behavior patterns with file:line citations." $(git log -1 --format="%H")
+
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 2 Sweep B complete"
+```
+
+### Task 2.3: Sweep C — Hidden notes → `llm_behavior_catalog.md` (Part 2 appended) (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md` (append hidden-notes section)
+
+- [ ] **Step 1: Build the Sweep C prompt**
+
+```
+You are a Tier 3 Worker sub-agent dispatched for a meta-tooling review track.
+
+TASK: Scan the following files for INLINE notes about LLM agent behavior, agent failure modes, anti-patterns, or workflow concerns that may not be captured in docs/reports/ or conductor/tracks/<id>/:
+
+1. C:/projects/manual_slop/AGENTS.md (root) — sections §Critical Anti-Patterns, §Session-Learned Anti-Patterns, §Process Anti-Patterns
+2. C:/projects/manual_slop/conductor/*.md (all .md files in conductor/ EXCEPT tracks/ subdirectory and code_styleguides/ subdirectory)
+3. C:/projects/manual_slop/docs/guide_*.md (36 files; the 14 deep-dive guides)
+4. C:/projects/manual_slop/conductor/code_styleguides/*.md (11 styleguide files)
+
+For each file, grep for: agent, anti-pattern, behavior, behavior drift, hard gate, hard ban, regression, heuristic, audit, slip, slop, scope creep, fix forward, surrender, escalation, tier-1, tier-2.
+
+For each match, capture: file:line + the surrounding 3 lines of context.
+
+OUTPUT FORMAT: Append to C:/projects/manual_slop/conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md
+
+Append structure:
+- H1 "Hidden Notes Appendix"
+- H2 "By Source File" (sub-sections per file)
+- Each file's section: bullet list of file:line + excerpt + which of the 12 LLM patterns it relates to (or "NEW finding not in the 12 patterns")
+
+CONSTRAINTS:
+- Total appendix: 200-300 LOC
+- Use 1-space indentation for nested content
+- No day estimates
+- If a source file has 0 matches, omit it
+
+Return the FULL APPENDIX content in your final message (not a summary). Do NOT rewrite the existing 12-pattern content; the Tier 1 orchestrator will append it.
+```
+
+- [ ] **Step 2: Dispatch Sweep C**
+
+Run:
+```bash
+uv run python scripts/mma_exec.py --role tier3-worker "<paste the prompt above>"
+```
+
+- [ ] **Step 3: Append output to `llm_behavior_catalog.md`**
+
+Read the sub-agent's output. Use `Get-Content` to load the existing `llm_behavior_catalog.md`, append the sub-agent's content, and write back.
+
+```powershell
+$existing = Get-Content -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620\llm_behavior_catalog.md" -Raw
+$appendix = "<sub-agent output>"
+Set-Content -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620\llm_behavior_catalog.md" -Value ($existing + "`n`n" + $appendix) -NoNewline
+```
+
+- [ ] **Step 4: Verify appended file**
+
+Run: `Get-ChildItem -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620\llm_behavior_catalog.md" | Select-Object Length`
+Expected: length > 22,000 bytes (700-1000 LOC total).
+
+- [ ] **Step 5: Spot-check 5 hidden-note citations**
+
+Open 5 random file:line citations from the appendix. Verify each exists.
+
+- [ ] **Step 6: Stage and commit + update state**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Sweep C (Tier 3 hidden notes appended to llm_behavior_catalog.md)"
+
+git notes add -m "Tier 3 Sweep C output: llm_behavior_catalog.md Part 2 (hidden notes appendix, ~200-300 LOC). Evidence from AGENTS.md + conductor/*.md + docs/guide_*.md + conductor/code_styleguides/*.md." $(git log -1 --format="%H")
+
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 2 (Sweeps A+B+C) complete"
+```
+
+- [ ] **Step 7: Bump `current_phase` to 3**
+
+Update `state.toml`:
+- `current_phase = 3`
+- `phase_2 = { status = "completed", checkpointsha = "<hash>", name = "Tier 3 sub-agent sweeps" }`
+
+Commit:
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 2 complete (all 3 sweeps done); advance to Phase 3"
+```
+
+---
+
+## Phase 3: Tier 1 Anchor Read (0 commits; internal scratchpad)
+
+**Scope:** Tier 1 reads 10 anchor reports to anchor the synthesis; produces an internal scratchpad (NOT committed; lives in conversation context).
+
+### Task 3.1: Read 10 anchor reports + write scratchpad (0 commits)
+
+- [ ] **Step 1: Read `docs/reports/CHRONOLOGY_TRACK_HANDOVER_20260620.md`**
+
+Read the full document. Capture in scratchpad:
+- The 5 "Lessons Learned" bullets (manual review bypass, metadata snapshot, git log as audit log, ask when in doubt, user said "manual review" twice)
+- The cleanup recommendations
+
+- [ ] **Step 2: Read `docs/reports/TRACK_COMPLETION_result_migration_review_pass_20260617.md`**
+
+Capture: the 43-site classification + 10 new heuristics + 3 audit-script bugs documented
+
+- [ ] **Step 3: Read `docs/reports/TRACK_COMPLETION_result_migration_small_files_20260617.md`**
+
+Capture: the Phase 10 REJECTION (5 laundering heuristics); Phase 11 redo; the 21 sites re-classified
+
+- [ ] **Step 4: Read `docs/reports/TRACK_COMPLETION_result_migration_app_controller_20260618.md`**
+
+Capture: the 6 phases + 45 sites + 22 stays
+
+- [ ] **Step 5: Read `docs/reports/TRACK_COMPLETION_result_migration_gui_2_20260619.md`**
+
+Capture: the 13 phases + 42 sites + anti-sliming protocol
+
+- [ ] **Step 6: Read `docs/reports/TRACK_COMPLETION_result_migration_baseline_cleanup_20260620.md`**
+
+Capture: the 3 last failures (broken `raise ErrorInfo from exc`, sentinel-None flagged UNCLEAR, lost `global` declarations)
+
+- [ ] **Step 7: Read `docs/reports/STATUS_REPORT_phase6_compact.md`**
+
+Capture: the unreachable code regression + the Lesson Learned ("NEVER extract a function with side effects and place the call AFTER a try/except that always returns")
+
+- [ ] **Step 8: Read `docs/reports/TIER1_REVIEW_phase9_dilemma_20260620.md`**
+
+Capture: the 4 options Tier 2 considered; the user's directive on Heuristic E; the meta-rule "Tier 2 cannot unilaterally add audit heuristics"
+
+- [ ] **Step 9: Read `docs/reports/EXCEPTION_HANDLING_AUDIT_20260616.md`**
+
+Capture: the 211 violations + 42 files + 5 doc gaps closed (the migration campaign's baseline)
+
+- [ ] **Step 10: Read `docs/reports/SESSION_REPORT_superpowers_review_init_20260619.md`**
+
+Capture: the user's framing "conservative changes incrementally to improve AI performance and quality standards of output"; the 16-section structure of that future report
+
+- [ ] **Step 11: Read `docs/reports/TRACK_COMPLETION_tier2_leak_prevention_20260620.md`**
+
+Capture: the 3-layer defense (permission + pre-commit hook + audit); the 6 design decisions
+
+- [ ] **Step 12: Write scratchpad summary**
+
+Produce a markdown summary in conversation context (NOT a file). The summary covers:
+- The 12 LLM behavior patterns (per spec §3.1 Part 2) — for each, the 3-5 anchor citations already identified
+- The 5-7 cross-cutting findings (what appears in multiple anchor reports)
+- The 3-5 "what AGENTS.md doesn't yet cover" candidates (for Part 2 §2.10)
+
+---
+
+## Phase 4: Part 1 Synthesis (1 commit)
+
+**Scope:** Write Part 1 of `report.md` (What Shipped; 5 sub-sections × 160-200 LOC each = 800-1000 LOC).
+
+### Task 4.1: Synthesize Part 1 from `shipped_work_index.md` (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/report.md` (populate §1.1 through §1.5)
+
+- [ ] **Step 1: Read `shipped_work_index.md`**
+
+Read the full file. This is the per-track scaffolding for Part 1.
+
+- [ ] **Step 2: Write §1.1 — The Result Migration Campaign**
+
+In `report.md`, replace `[160-200 LOC; populated in Phase 4]` under `### §1.1 The Result Migration Campaign (5 sub-tracks + umbrella)` with:
+
+```
+The Result Migration campaign (umbrella `result_migration_20260616` + 5 sub-tracks) shipped 2026-06-17 → 2026-06-20. All 5 sub-tracks landed with 100% `Result[T]` convention coverage across the modified files.
+
+[160-200 LOC covering:
+- Umbrella scope (5 sub-tracks; 268 sites per audit)
+- Sub-track 1: result_migration_review_pass_20260617 (43 sites classified; 10 heuristics added)
+- Sub-track 2: result_migration_small_files_20260617 (Phase 10 REJECTED → Phase 11 redo → Phase 12 migration; 5 laundering heuristics reverted)
+- Sub-track 3: result_migration_app_controller_20260618 (45 sites; 6 phases; 22 stays)
+- Sub-track 4: result_migration_gui_2_20260619 (42 sites; 13 phases; anti-sliming protocol formalized)
+- Sub-track 5: result_migration_baseline_cleanup_20260620 (88 sites across 3 files: mcp_client 46 + ai_client 33 + rag_engine 9)
+- Key reports cited (file:line): RESULT_MIGRATION_CAMPAIGN_STATUS_20260619.md, the 5 TRACK_COMPLETION_*.md, PHASE12_TRIAGE_20260617.md, TIER1_REVIEW_phase9_dilemma_20260620.md
+- LLM-behavior touch-points: §2.1 anti-sliming, §2.3 regression-after-refactor, §2.4 heuristic proliferation, §2.8 lost-context extraction]
+```
+
+- [ ] **Step 3: Write §1.2 — Tier 2 Autonomous Sandbox Family**
+
+Replace the §1.2 placeholder with [160-200 LOC covering: the 4 sandbox tracks; the 3-layer defense in leak prevention; the bypass-and-document-deviation pattern in chronology handover].
+
+- [ ] **Step 4: Write §1.3 — Stability & Test-Infrastructure Tracks**
+
+Replace the §1.3 placeholder with [160-200 LOC covering: public_api_migration_and_ui_polish_20260615 + rag_test_failures_20260615 + live_gui_test_fixes_20260618 + test_sandbox_hardening_20260619 + exception_handling_audit_20260616].
+
+- [ ] **Step 5: Write §1.4 — Meta-Analysis Corpus**
+
+Replace the §1.4 placeholder with [160-200 LOC covering: nagent_review v3.1 (2,214 lines) + superpowers_review_init (blocked by chronology) + fable_review + intent_dsl_survey + chronology_20260619 (Phase 10 pending sign-off)].
+
+- [ ] **Step 6: Write §1.5 — One-Off Fixes & Polishes**
+
+Replace the §1.5 placeholder with [160-200 LOC covering: ai_loop_regressions_20260614 + doeh_test_thinking_cleanup_20260615 + send_result_to_send_20260616 + ai_client_docs_20260613 + ai_decoupling_revert].
+
+- [ ] **Step 7: Self-review Part 1**
+
+Verify:
+- Each sub-section is 160-200 LOC (count with `wc -l` or read line count)
+- Each sub-section has file:line citations
+- Total Part 1 LOC: 800-1000
+
+If any sub-section is under 150 LOC, expand with more detail from the `shipped_work_index.md`.
+
+- [ ] **Step 8: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/report.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 4 Part 1 (What Shipped, 5 sub-sections)"
+```
+
+- [ ] **Step 9: Attach git note + commit state**
+
+```bash
+git notes add -m "Part 1 of report.md: What Shipped. 5 track-family sub-sections (~160-200 LOC each, 800-1000 LOC total). Uses shipped_work_index.md (Tier 3 sweep A) as scaffolding. File:line citations throughout." $(git log -1 --format="%H")
+
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 4 (Part 1) complete"
+```
+
+---
+
+## Phase 5: Part 2 Synthesis (1-2 commits)
+
+**Scope:** Write Part 2 of `report.md` (LLM Behavior Patterns; 12 patterns × 125-170 LOC each = 1500-2000 LOC). Split into 2 commits if LOC > 1500 (commits at §2.6 and §2.12).
+
+### Task 5.1: Synthesize Part 2 from `llm_behavior_catalog.md` (1 commit, possibly +1)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/report.md` (populate §2.1 through §2.12)
+
+- [ ] **Step 1: Read `llm_behavior_catalog.md`**
+
+Read the full file. This is the evidence scaffolding for Part 2.
+
+- [ ] **Step 2: Verify 12 patterns have sufficient evidence**
+
+For each pattern in the catalog, check that there are ≥3 evidence citations. If any pattern has <3, mark it as "INSUFFICIENT EVIDENCE" in the report (1 sentence + drop from Part 2 if necessary).
+
+If the actual count is <8, drop patterns until ≥8 remain. Update `state.toml` `t5_1_part2_synthesis.description` with the actual count.
+
+- [ ] **Step 3: Write §2.1 — Anti-Sliming**
+
+Replace the §2.1 placeholder with the 7-sub-section structure per spec §3.1 Part 2:
+- §N.1 What N adds (1 sentence)
+- §N.2 Driver/structure (what causes the pattern)
+- §N.3 Invariants (what should always hold)
+- §N.4 Per-commit detail (3-5 file:line citations with excerpts from llm_behavior_catalog.md)
+- §N.5 Manual Slop implications (2-3 paragraphs)
+- §N.6 Honest gaps (≥6 bullets)
+- §N.7 Code-shape sketch (1 paragraph with `{ssdl}` tags)
+- Verdict block (NEW / PARTIALLY-CODIFIED / FULLY-CODIFIED / SUBSUMED)
+
+- [ ] **Step 4: Write §2.2 — Hard-Gate Bypass**
+
+Same 7-sub-section structure. Citations from `CHRONOLOGY_TRACK_HANDOVER_20260620.md` §"Lessons learned" #1.
+
+- [ ] **Step 5: Write §2.3 — Regression-After-Refactor**
+
+Same structure. Citations from `STATUS_REPORT_phase6_compact.md` §2 + `TRACK_COMPLETION_result_migration_baseline_cleanup_20260620.md` §4 Failure 3 + AGENTS.md §"Indentation-Driven Class Method Visibility".
+
+- [ ] **Step 6: Write §2.4 — Heuristic Proliferation Mid-Track**
+
+Same structure. Citations from `TIER1_REVIEW_phase9_dilemma_20260620.md` + `TRACK_COMPLETION_result_migration_baseline_cleanup_20260620.md` Phase 9.
+
+- [ ] **Step 7: Write §2.5 — Tier 2 Escalation Drift**
+
+Same structure. Citations from `CHRONOLOGY_TRACK_HANDOVER_20260620.md` §"Lessons learned" #5.
+
+- [ ] **Step 8: Write §2.6 — Report-As-Substitute-For-Fix**
+
+Same structure. Citations from `CHRONOLOGY_TRACK_HANDOVER_20260620.md` (the entire document) + AGENTS.md §"Process Anti-Patterns" #2.
+
+- [ ] **Step 9: Write §2.7 — Decision-Deflection**
+
+Same structure. Citations from `docs/reports/*.md` "next steps" sections + AGENTS.md §"Process Anti-Patterns" #6.
+
+- [ ] **Step 10: Write §2.8 — Lost-Context Extraction**
+
+Same structure. Citations from §2.3 evidence + AGENTS.md §"Session-Learned Anti-Patterns" #2.
+
+- [ ] **Step 11: Write §2.9 — Literal-vs-Inferred Instruction Interpretation**
+
+Same structure. Citations from `CHRONOLOGY_TRACK_HANDOVER_20260620.md` §"Lessons learned" #5 + AGENTS.md §"Session-Learned Anti-Patterns" #4.
+
+- [ ] **Step 12: Write §2.10 — Cross-Track Synthesis Gap**
+
+Same structure. Citations from the synthesis itself (the patterns NOT yet in AGENTS.md).
+
+- [ ] **Step 13: Write §2.11 — "I'm Done" Surrender Threshold**
+
+Same structure. Citations from AGENTS.md §"Process Anti-Patterns" #6 + #8 + `STATUS_REPORT_phase6_compact.md`.
+
+- [ ] **Step 14: Write §2.12 — Anti-Sliming Protocol Evolution**
+
+Same structure. Citations from `TRACK_COMPLETION_result_migration_baseline_cleanup_20260620.md` Phase 10-13 + `TIER1_REVIEW_phase9_dilemma_20260620.md`.
+
+- [ ] **Step 15: Self-review Part 2**
+
+Verify per spec §3.9 chunking strategy:
+- Each pattern: ≥4 sub-sections (target: 7), ≥3 file:line citations, ≥2 honest gaps, ≥1 Manual Slop implication paragraph, verdict block present
+- Total Part 2 LOC: 1500-2000
+
+If any pattern is under 100 LOC, expand. If total < 1500, add more detail to existing patterns.
+
+- [ ] **Step 16: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/report.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 5 Part 2 (LLM Behavior Patterns, 12 patterns)"
+```
+
+If Part 2 LOC > 1500, commit at §2.6 and again at §2.12 (2 commits instead of 1).
+
+- [ ] **Step 17: Attach git note + commit state**
+
+```bash
+git notes add -m "Part 2 of report.md: LLM Behavior Patterns. 12 patterns (~125-170 LOC each, 1500-2000 LOC total) with the 7-sub-section structure + verdict block. Uses llm_behavior_catalog.md (Tier 3 sweeps B+C) as evidence scaffolding." $(git log -1 --format="%H")
+
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 5 (Part 2) complete"
+```
+
+---
+
+## Phase 6: Part 3 Synthesis (1 commit)
+
+**Scope:** Write Part 3 of `report.md` (Workflow Improvements; 15-25 improvements × 50-80 LOC each = 1000-1200 LOC).
+
+### Task 6.1: Synthesize Part 3 from Part 2 patterns + Tier 1 judgment (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/report.md` (populate §AGENTS.md through §scripts/audit_*.py)
+
+- [ ] **Step 1: Derive the improvement list from Part 2**
+
+For each of the 12 patterns in Part 2, identify:
+- Is there a workflow improvement that addresses it? (Most patterns → 1 improvement; some → 2)
+- Which target doc does it go to? (AGENTS.md / conductor/workflow.md / conductor/code_styleguides/error_handling.md / .opencode/agents/tier2-autonomous.md / scripts/audit_*.py)
+- Which confidence tier? (Tier 1 = apply now / Tier 2 = defer 1 cycle / Tier 3 = open question)
+
+Produce an internal scratchpad (NOT committed) of the 15-25 improvements.
+
+- [ ] **Step 2: Write §AGENTS.md section**
+
+Replace the placeholder with:
+- H3 sub-section per improvement targeting AGENTS.md (likely 3-5 entries)
+- Each improvement: H4 title + the 6-sub-section structure per spec §3.1 Part 3:
+  - What (1-sentence change)
+  - Why (evidence from Part 2 with file:line citations)
+  - Where (file:line destination)
+  - Risk (what could break)
+  - Verification (how the user checks)
+  - Rollback (how to revert)
+
+- [ ] **Step 3: Write §conductor/workflow.md section**
+
+Replace with H3 sub-sections for each workflow.md improvement (likely 2-3 Tier 1 + 1-2 Tier 2 + 0-1 Tier 3).
+
+- [ ] **Step 4: Write §conductor/code_styleguides/error_handling.md section**
+
+Replace with H3 sub-sections for each styleguide improvement (likely 1-2 Tier 1 + 1 Tier 2).
+
+- [ ] **Step 5: Write §.opencode/agents/tier2-autonomous.md section**
+
+Replace with H3 sub-sections for each .opencode improvement (likely 1-2 Tier 1 + 0-1 Tier 2 + 1 Tier 3).
+
+- [ ] **Step 6: Write §scripts/audit_*.py section**
+
+Replace with H3 sub-sections for each audit-script improvement (likely 0-1 Tier 1 + 2-3 Tier 2 + 1 Tier 3).
+
+- [ ] **Step 7: Self-review Part 3**
+
+Verify:
+- Total improvements: 15-25
+- Each improvement has all 6 sub-sections (What / Why / Where / Risk / Verification / Rollback)
+- Each improvement has a destination file path
+- Each improvement has a risk tier
+- Total Part 3 LOC: 1000-1200
+
+- [ ] **Step 8: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/report.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 6 Part 3 (Workflow Improvements, 15-25 entries by target doc × confidence tier)"
+```
+
+- [ ] **Step 9: Attach git note + commit state**
+
+```bash
+git notes add -m "Part 3 of report.md: Workflow Improvements. 15-25 improvements (1000-1200 LOC) by 5 target docs x 3 confidence tiers. Each with the 6-sub-section structure (What/Why/Where/Risk/Verification/Rollback). Ready for the rebuild track to read." $(git log -1 --format="%H")
+
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 6 (Part 3) complete"
+```
+
+---
+
+## Phase 7: Part 4 Synthesis (1 commit)
+
+**Scope:** Write Part 4 of `report.md` (Implementation Sequencing; 5 phases × 60-100 LOC each = 300-500 LOC).
+
+### Task 7.1: Synthesize Part 4 from Part 3 improvements (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/report.md` (populate §Phase 1 through §Phase 5)
+
+- [ ] **Step 1: Group Part 3 improvements by sequencing phase**
+
+For each Part 3 improvement, assign a phase (1-5) per spec §3.1 Part 4:
+- Phase 1: AGENTS.md doc edits
+- Phase 2: conductor/workflow.md process gates
+- Phase 3: conductor/code_styleguides/error_handling.md convention tightening
+- Phase 4: .opencode/ tier-specific directives
+- Phase 5: scripts/audit_*.py + CI gate additions
+
+Each improvement should appear in exactly one phase. Phase ordering is conservative: zero-risk first, audit scripts last.
+
+- [ ] **Step 2: Write §Phase 1 — AGENTS.md Doc Edits**
+
+Replace the placeholder with:
+- §N.1 Scope (which improvements from Part 3 §AGENTS.md)
+- §N.2 Risk assessment (zero; doc-only)
+- §N.3 Verification (the user reads the diff)
+- §N.4 Rollback path (`git revert <commit-sha>`)
+- §N.5 Open questions (e.g., "Does the user want to bump the anti-pattern number, or add a new section?")
+
+- [ ] **Step 3: Write §Phase 2 — conductor/workflow.md Process Gates**
+
+Same 5-sub-section structure.
+
+- [ ] **Step 4: Write §Phase 3 — conductor/code_styleguides/error_handling.md**
+
+Same structure.
+
+- [ ] **Step 5: Write §Phase 4 — .opencode/ Tier-Specific Directives**
+
+Same structure. Risk assessment: Medium (changes how Tier 2 interprets instructions).
+
+- [ ] **Step 6: Write §Phase 5 — scripts/audit_*.py + CI Gate Additions**
+
+Same structure. Risk assessment: Medium-High (audit script is enforcement).
+
+- [ ] **Step 7: Self-review Part 4**
+
+Verify:
+- Each phase has all 5 sub-sections
+- Each phase has at least 1 rollback command
+- Total Part 4 LOC: 300-500
+- Phase ordering is conservative (low-risk first)
+
+- [ ] **Step 8: Verify total report.md LOC**
+
+Run: `Get-ChildItem -LiteralPath "conductor\tracks\meta_tooling_workflow_review_20260620\report.md" | Select-Object Length`
+Divide by ~30 bytes/LOC for markdown.
+Expected: ≥4,000 LOC.
+
+If < 4,000, expand the thinnest sub-section in Part 1, 2, or 3.
+
+- [ ] **Step 9: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/report.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 7 Part 4 (Implementation Sequencing, 5 phases)"
+```
+
+- [ ] **Step 10: Attach git note + commit state**
+
+```bash
+git notes add -m "Part 4 of report.md: Implementation Sequencing. 5 conservative phases (60-100 LOC each, 300-500 LOC total). Each with Scope/Risk/Verification/Rollback/OpenQuestions. Total report.md >= 4,000 LOC target met." $(git log -1 --format="%H")
+
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 7 (Part 4) complete; report.md 4K floor met"
+```
+
+---
+
+## Phase 8: Side Artifacts + Standalone Inputs (5 commits)
+
+**Scope:** Write the 5 side artifacts that complement `report.md` and the 2 standalone inputs for the rebuild track.
+
+### Task 8.1: Write `comparison_table.md` (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md`
+
+- [ ] **Step 1: Extract ~50 tracks from `shipped_work_index.md`**
+
+Read `shipped_work_index.md`. For each track family, list ~10 tracks (5 families × 10 = 50).
+
+- [ ] **Step 2: Populate the table**
+
+Replace the placeholder with the populated table per spec §3.2:
+| Track family | Track name | Status | Key reports | First LLM-behavior tag |
+
+For each row:
+- Track family: one of migration campaign / tier-2 sandbox / stability-test-infra / meta-analysis / one-off polish
+- Track name: the track folder name
+- Status: Shipped / In flight / Pending sign-off / Abandoned / Superseded
+- Key reports: 1-3 file names
+- First LLM-behavior tag: Part 2 §N.N number (e.g., "2.3" for Phase 6 unreachable-code regression)
+
+- [ ] **Step 3: Commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 comparison_table.md (~50 rows)"
+git notes add -m "~50-row flat reference of past-month tracks. Columns: track family / track name / status / key reports / first LLM-behavior tag." $(git log -1 --format="%H")
+```
+
+### Task 8.2: Write `decisions.md` (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md`
+
+- [ ] **Step 1: Extract improvements from Part 3**
+
+For each of the 15-25 improvements in Part 3, create a decisions.md row per spec §3.3.
+
+- [ ] **Step 2: Sort by priority**
+
+Sort: HIGH (Phase 1-2 improvements) → MEDIUM (Phase 3) → LOW (Phase 4-5).
+
+- [ ] **Step 3: Populate the table**
+
+Replace the placeholder with the populated table. Include all 8 fields per spec §3.3.
+
+- [ ] **Step 4: Commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 decisions.md (15-25 entries)"
+git notes add -m "Prioritized rebuild backlog (15-25 entries). Sorted HIGH to LOW. Each entry has destination file + evidence + risk + sequencing phase." $(git log -1 --format="%H")
+```
+
+### Task 8.3: Write `nagent_takeaways_meta_tooling_20260620.md` (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md`
+
+- [ ] **Step 1: Read sibling review bridge artifacts**
+
+Read:
+- `conductor/tracks/nagent_review_20260608/nagent_takeaways_20260608.md`
+- `conductor/tracks/superpowers_review_20260619/nagent_takeaways_superpowers_20260619.md` (if exists)
+- `conductor/tracks/nagent_review_20260608/nagent_takeaways_v3_20260619.md`
+- `conductor/tracks/nagent_review_20260608/nagent_takeaways_v3_1_20260620.md`
+
+- [ ] **Step 2: Write TL;DR (1 paragraph)**
+
+In `nagent_takeaways_meta_tooling_20260620.md`, replace the TL;DR placeholder with a paragraph summarizing how this track's 12 LLM patterns relate to the 4 sibling reviews' verdicts.
+
+- [ ] **Step 3: Write Cross-Reference Table**
+
+Replace the table placeholder with ~10-15 rows: one per LLM pattern that touches a verdict in a sibling review.
+
+- [ ] **Step 4: Write The N New Findings section**
+
+Replace with paragraphs for each new finding (likely 4-6: anti-sliming protocol, Phase 9 dilemma, chronology handover pattern, regression-after-refactor, the audit-script-as-convention issue).
+
+- [ ] **Step 5: Write Contradictions/Extensions section**
+
+Replace with paragraphs for any sibling-review findings this track contradicts or extends.
+
+- [ ] **Step 6: Write Pointer to fable_review**
+
+Replace with 1 paragraph.
+
+- [ ] **Step 7: Commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 nagent_takeaways bridge (5-part structure)"
+git notes add -m "Bridge to the 4 sibling meta-analysis reviews. 5-part structure: TL;DR + cross-reference table + new findings + contradictions + fable pointer. ~200 LOC." $(git log -1 --format="%H")
+```
+
+### Task 8.4: Write `workflow_improvements.md` (standalone) (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md`
+
+- [ ] **Step 1: Copy Part 3 content**
+
+Copy the entire Part 3 (§AGENTS.md through §scripts/audit_*.py) from `report.md`.
+
+- [ ] **Step 2: Strip cross-references to Part 1/2**
+
+For each improvement, remove the "(see Part 2 §2.X)" cross-references; replace with direct file:line citations to the source reports.
+
+- [ ] **Step 3: Add a "How to use this document" header**
+
+Add a 1-paragraph preamble: "This is the standalone catalog of workflow improvements the rebuild track should read. The full 4,000-LOC report.md contains the supporting evidence; this document assumes you've decided to apply the changes."
+
+- [ ] **Step 4: Commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 workflow_improvements.md (standalone Part 3)"
+git notes add -m "Standalone Part 3 input for the rebuild track. Read this WITHOUT re-reading the 4,000-LOC report.md. Verbatim copy of Part 3 with cross-references stripped." $(git log -1 --format="%H")
+```
+
+### Task 8.5: Write `implementation_sequencing.md` (standalone) (1 commit)
+
+**Files:**
+- Modify: `conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md`
+
+- [ ] **Step 1: Copy Part 4 content**
+
+Copy the entire Part 4 (§Phase 1 through §Phase 5) from `report.md`.
+
+- [ ] **Step 2: Add Phase Dependencies section**
+
+Append a new H1 section "Phase Dependencies" with a table:
+| Phase | Depends on | Blocks |
+|---|---|---|
+| 1 | (nothing) | 2, 3, 4, 5 |
+| 2 | 1 | 3, 4 |
+| 3 | 2 | 4 |
+| 4 | 3 | 5 |
+| 5 | 4 | (nothing) |
+
+- [ ] **Step 3: Add a "How to use this document" header**
+
+Add a 1-paragraph preamble.
+
+- [ ] **Step 4: Commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 implementation_sequencing.md (standalone Part 4 + phase dependencies)"
+git notes add -m "Standalone Part 4 input for the rebuild track. Verbatim copy of Part 4 + phase dependencies table. The 5-phase conservative ordering enforces zero-risk first, audit scripts last." $(git log -1 --format="%H")
+```
+
+---
+
+## Phase 9: Self-Review (0-1 commits)
+
+**Scope:** Per the brainstorming skill: placeholder scan, internal consistency, scope check, ambiguity check, chunking verification.
+
+### Task 9.1: Self-review pass (0-1 commits)
+
+- [ ] **Step 1: Placeholder scan**
+
+Search all 11 files for "TBD", "TODO", "fill in", "populated in Phase", "FIXME", "XXX". Fix inline.
+
+- [ ] **Step 2: Internal consistency check**
+
+Verify:
+- `comparison_table.md` rows match `report.md` Part 1 §N.N sub-sections
+- `decisions.md` entries match `report.md` Part 3 §X.Y.Z sub-sections
+- `workflow_improvements.md` is verbatim Part 3 (minus cross-refs)
+- `implementation_sequencing.md` is verbatim Part 4 (plus phase dependencies)
+- `nagent_takeaways_*.md` cross-references match the actual sibling review section numbers
+
+- [ ] **Step 3: Scope check**
+
+Verify:
+- No `src/` / `tests/` / `conductor/*.md` / `AGENTS.md` / `.opencode/` / `scripts/audit_*.py` changes
+- All recommendations go to `workflow_improvements.md` (not to source files)
+
+- [ ] **Step 4: Ambiguity check**
+
+Verify:
+- Every Part 3 improvement has a destination file path
+- Every Part 4 phase has a rollback command
+- Every Part 2 pattern has a verdict block
+
+- [ ] **Step 5: Chunking verification (per nagent_review_v3.1)**
+
+Verify per spec §3.9:
+- Each Part 2 pattern: ≥4 sub-sections (target 7), ≥3 file:line citations, ≥2 honest gaps, ≥1 Manual Slop implication, verdict block
+- Each Part 3 improvement: ≥4 sub-sections (target 6), ≥1 evidence citation, ≥1 verification step
+- Each Part 4 phase: ≥3 sub-sections (target 5), ≥1 rollback command
+
+If any check fails, fix inline.
+
+- [ ] **Step 6: Total report.md LOC check**
+
+Verify: total report.md LOC ≥ 4,000. If < 4,000, thicken the thinnest section.
+
+- [ ] **Step 7: Commit fixes (if any)**
+
+If any fixes were needed in Steps 1-6:
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 9 self-review fixes"
+git notes add -m "Self-review pass: placeholder scan + internal consistency + scope check + ambiguity check + chunking verification. <N> fixes applied." $(git log -1 --format="%H")
+```
+
+If no fixes needed, no commit. Update `state.toml` `t9_1_self_review.status = "completed"` with note "no fixes needed".
+
+---
+
+## Phase 10: User Review Gate (0 commits; user-driven)
+
+**Scope:** User reviews the final report + side artifacts + standalone inputs; approves or iterates.
+
+### Task 10.1: User review (0 commits)
+
+- [ ] **Step 1: Inform the user**
+
+Send a message to the user:
+> "Track artifacts complete. Please review:
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/report.md` (≥4,000 LOC; 4 parts)
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/comparison_table.md` (~50 rows)
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/decisions.md` (15-25 entries)
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/shipped_work_index.md`
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/llm_behavior_catalog.md`
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/nagent_takeaways_meta_tooling_20260620.md`
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/workflow_improvements.md` (standalone for rebuild track)
+> - `conductor/tracks/meta_tooling_workflow_review_20260620/implementation_sequencing.md` (standalone for rebuild track)
+>
+> Let me know if you want changes before I run Phase 11 (Finalize)."
+
+- [ ] **Step 2: Wait for user response**
+
+Do not proceed without explicit user approval.
+
+- [ ] **Step 3: Apply any user-requested changes**
+
+If the user requests changes, apply them as additional commits. Update `state.toml` per change.
+
+- [ ] **Step 4: Update `state.toml` to mark user review complete**
+
+After user approval:
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 10 (user review gate) complete (approved YYYY-MM-DD)"
+```
+
+---
+
+## Phase 11: Finalize (1 commit)
+
+**Scope:** Update `state.toml`, `metadata.json`, and `conductor/tracks.md` to reflect completion.
+
+### Task 11.1: Finalize (1 commit)
+
+- [ ] **Step 1: Update `state.toml`**
+
+Update:
+- `status = "active"` (per spec — track stays active until archived per the chronology archive convention)
+- `current_phase = 11`
+- All `phase_N.status = "completed"`
+- All `phase_N.checkpointsha = "<hash>"` (use the final commit of each phase)
+- All `tasks.status = "completed"`
+- `last_updated = "2026-06-20"` (or current date if different)
+- All `verification.* = true`
+
+- [ ] **Step 2: Update `metadata.json`**
+
+Update:
+- `status: "active"`
+- Add a new top-level field `"final_statistics"` with the actual counts:
+```json
+{
+  "report_total_loc": <actual>,
+  "pattern_count": <actual>,
+  "improvement_count": <actual>,
+  "phase_count": <actual>,
+  "commit_count": <actual>,
+  "sweep_a_loc": <actual>,
+  "sweep_b_loc": <actual>,
+  "sweep_c_loc": <actual>,
+  "comparison_table_rows": <actual>,
+  "decisions_entries": <actual>,
+  "shipped_work_index_tracks": <actual>
+}
+```
+
+- [ ] **Step 3: Update `conductor/tracks.md`**
+
+Move the track entry from "Active Tracks" to "Recently Completed" section (or whichever section the project's convention uses — check the existing format). Add a one-line summary:
+
+```
+- [x] **Track: Meta-Tooling Workflow Review — Past-Month LLM Behavior Analysis** `[checkpoint: <sha-7>]`  
+  *Link: [./tracks/meta_tooling_workflow_review_20260620/](./tracks/meta_tooling_workflow_review_20260620/), Spec + Plan: see track folder.  
+  *Goal: Analyze all past-month reports + git log + state.toml + guide docs; identify 8-16 LLM behavior patterns; catalog workflow improvements by target doc × confidence tier; sequence for a conservative rebuild. 4-part report ≥4,000 LOC. Research-only.*
+```
+
+- [ ] **Step 4: Stage and commit**
+
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml conductor/tracks/meta_tooling_workflow_review_20260620/metadata.json conductor/tracks.md
+git commit -m "conductor(track): meta_tooling_workflow_review_20260620 Phase 11 finalize"
+```
+
+- [ ] **Step 5: Attach git note**
+
+```bash
+git notes add -m "Track complete. state.toml advanced to current_phase=11. metadata.json updated with final statistics (report LOC: <N>, patterns: <N>, improvements: <N>). conductor/tracks.md updated. Ready for the near-future workflow-improvements rebuild track to consume workflow_improvements.md + implementation_sequencing.md as standalone inputs." $(git log -1 --format="%H")
+```
+
+- [ ] **Step 6: Mark Phase 11 task complete**
+
+Update `state.toml`:
+- `t11_1_finalize.status = "completed"`
+- `t11_1_finalize.commit_sha = "<hash>"`
+
+Commit:
+```bash
+git add conductor/tracks/meta_tooling_workflow_review_20260620/state.toml
+git commit -m "conductor(plan): mark Phase 11 (Finalize) complete; meta_tooling_workflow_review_20260620 SHIPPED"
+```
+
+---
+
+## Plan Self-Review
+
+**1. Spec coverage:**
+- §3.1 Part 1 (5 sub-sections) → Phase 4 ✓
+- §3.1 Part 2 (12 patterns) → Phase 5 ✓
+- §3.1 Part 3 (15-25 improvements) → Phase 6 ✓
+- §3.1 Part 4 (5 phases) → Phase 7 ✓
+- §3.2 comparison_table.md → Task 8.1 ✓
+- §3.3 decisions.md → Task 8.2 ✓
+- §3.4 shipped_work_index.md → Task 2.1 ✓
+- §3.5 llm_behavior_catalog.md → Tasks 2.2 + 2.3 ✓
+- §3.6 nagent_takeaways_meta_tooling_20260620.md → Task 8.3 ✓
+- §3.7 workflow_improvements.md (standalone) → Task 8.4 ✓
+- §3.8 implementation_sequencing.md (standalone) → Task 8.5 ✓
+- §3.9 chunking strategy (≥4K LOC) → Phase 9 self-review verifies ✓
+
+**2. Placeholder scan:** No "TBD" / "TODO" / "fill in details" in the plan. Each task has explicit file paths + content + commands.
+
+**3. Type consistency:** N/A (research-only track; no code changes; no type definitions).
+
+**4. Ambiguity check:**
+- Each Tier 3 sub-agent prompt has explicit file paths + output format + LOC target ✓
+- Each Phase 4-7 task has explicit LOC range per sub-section ✓
+- The ≥4,000 LOC report floor is checked in Phase 9 ✓
+
+**Minor adjustments applied during self-review:**
+- Phase 6 (Part 3) added explicit "internal scratchpad" instruction in Step 1 (Tier 1 judgment call; not committed)
+- Phase 7 Step 8 added explicit LOC check (the ≥4,000 floor verification)
+- Phase 8 Tasks 8.4 + 8.5 added explicit "strip cross-references" / "add preamble" instructions for the standalone docs
+
+Plan ready for execution.
+
+---
+
+## Execution Handoff
+
+**Plan complete and saved to `conductor/tracks/meta_tooling_workflow_review_20260620/plan.md`. Two execution options:**
+
+**1. Subagent-Driven (recommended)** — Tier 2 Tech Lead dispatches a fresh Tier 3 sub-agent per task (well-suited for the Tier 3 sweep dispatches in Phase 2; less ideal for the Tier 1 synthesis phases 3-7 which require sustained context). The Tier 2 review between tasks catches the anti-sliming risk in the synthesis phases.
+
+**2. Inline Execution** — Tier 1 Orchestrator executes tasks in this session using executing-plans. The Tier 3 sweeps (Phase 2) use `scripts/mma_exec.py --role tier3-worker`; the Tier 1 synthesis phases (3-7) run inline with full context. Total session token burn: substantial (the 4,000-LOC report is generated in-conversation).
+
+**My recommendation: Option 2 (inline Tier 1 execution).** The 4-part report synthesis requires sustained Tier 1 context; splitting it across sub-agents risks losing the cross-cutting synthesis voice. Phase 2 (Tier 3 sweeps) is the only sub-agent dispatch. Phases 8-11 are mechanical and can run inline.
+
+**Which approach do you want?**
\ No newline at end of file