docs(M00): milestone closeout, invariant registry, ledger update

- Add Invariant Registry and Serena Refactor Principles to serena.md
- Create M00_run1.md, M00_summary.md, M00_audit.md
- Update milestone ledger: M00 Completed
- Document CI evidence (Linter pass; Tests fail pre-existing)

Made-with: Cursor

docs/milestones/M00/M00_audit.md

@@ -0,0 +1,88 @@
+# M00 Milestone Audit
+
+**Date:** 2025-03-07  
+**Milestone:** M00 — Program Kickoff, Baseline Freeze, Phase Map, and E2E Verification  
+**Auditor:** Serena governance (self-audit)
+
+---
+
+## 1. Executive Summary
+
+M00 successfully established the Serena refactor program baseline and governance structure. All deliverables are documentation and verification artifacts. No runtime code was modified. Baseline integrity verified. Invariant registry and principles added to the ledger.
+
+**Audit score: 5 / 5**
+
+---
+
+## 2. Invariant Verification
+
+| Invariant | Verified |
+|-----------|----------|
+| Existing application startup behavior unchanged | ✓ No code changes |
+| Existing API routes and UI entry unchanged | ✓ No code changes |
+| Existing extension discovery/load unchanged | ✓ No code changes |
+| Existing test semantics unchanged | ✓ No test changes |
+| No CI weakening | ✓ No workflow changes |
+| No new runtime dependencies in hot paths | ✓ None added |
+| No structural refactor | ✓ Docs and scripts only |
+| docs-only changes behavior-preserving | ✓ Scripts wrap existing commands |
+
+---
+
+## 3. Documentation Completeness
+
+| Artifact | Status |
+|----------|--------|
+| docs/serena.md | ✓ Ledger, phase map, invariants, registry, principles |
+| M00_plan.md | ✓ Scope, invariants, acceptance criteria |
+| M00_preflight.md | ✓ Surfaces, constraints, environment |
+| M00_e2e_baseline.md | ✓ Baseline freeze, verification commands |
+| M00_ci_inventory.md | ✓ Workflows, gaps |
+| M00_toolcalls.md | ✓ Tool log |
+| M00_run1.md | ✓ CI analysis |
+| M00_summary.md | ✓ Outcomes, baseline ref |
+| M00_audit.md | ✓ This document |
+
+---
+
+## 4. CI Signal Integrity
+
+| Job | Result | Notes |
+|-----|--------|-------|
+| Linter | PASS | Ruff and eslint pass on M00 branch |
+| Tests | FAIL | Pre-existing: CLIP install fails (pkg_resources). Not from M00. |
+
+**Assessment:** Linter validates M00 additions. Test failure is a known fork CI environment issue; M01 will address.
+
+---
+
+## 5. Risk Assessment
+
+| Risk | Mitigation |
+|------|------------|
+| Baseline tag moved | Tag is annotated and pushed; documented in ledger |
+| Documentation drift | serena.md is source of truth; ledger updated |
+| Test failure blocks merge | Documented as pre-existing; M00 is docs-only |
+
+---
+
+## 6. Repository Cleanliness
+
+| Check | Result | Notes |
+|-------|--------|-------|
+| ruff | PASS | CI uses ruff 0.3.3; passes on M00 branch |
+| pytest | Not run | Requires server startup (launch.py --test-server); documented in M00_e2e_baseline.md |
+| mypy | Not configured | No mypy in repo or CI; not in scope for M00 |
+
+---
+
+## 7. Audit Score
+
+**5 / 5**
+
+**Rationale:**
+- No runtime change
+- Baseline freeze verified (tag → 82a973c0)
+- Ledger created with phase map, invariants, registry, principles
+- All M00 artifacts present and consistent
+- CI failure is pre-existing and documented truthfully

docs/milestones/M00/M00_run1.md

@@ -0,0 +1,85 @@
+# M00 CI Run 1 — Workflow Analysis
+
+**Date:** 2025-03-07  
+**Branch:** m00-kickoff-baseline-e2e  
+**Commit:** 0a8ade1a (M00: Program kickoff, baseline freeze, phase map, E2E verification)
+
+---
+
+## 1. Workflow Identity
+
+| Workflow | Run ID | Trigger | Status |
+|----------|--------|---------|--------|
+| Linter | 22790940335 | push | success |
+| Tests | 22790940333 | push | failure |
+
+**Repo:** m-cahill/serena
+
+---
+
+## 2. Change Context
+
+M00 introduces **documentation and governance only**:
+- docs/serena.md
+- docs/milestones/M00/* (plan, preflight, e2e_baseline, ci_inventory, toolcalls)
+- scripts/dev/run_m00_baseline_e2e.ps1, .sh
+
+**No runtime code changes.** No modifications to modules/, launch.py, webui.py, or workflows.
+
+---
+
+## 3. Job Inventory
+
+### Linter (22790940335)
+
+| Job | Result | Duration |
+|-----|--------|----------|
+| ruff | success | ~18s |
+| eslint | success | (included) |
+
+**Signal integrity:** PASS. All lint checks passed.
+
+### Tests (22790940333)
+
+| Job | Result | Duration |
+|-----|--------|----------|
+| tests on CPU with empty model | failure | ~42s |
+
+**Failure cause:** Pre-existing CI environment issue. `launch.py --skip-torch-cuda-test --exit` (Setup environment step) fails during CLIP installation:
+
+```
+ModuleNotFoundError: No module named 'pkg_resources'
+RuntimeError: Couldn't install clip.
+```
+
+This is a dependency/build environment issue (setuptools/pkg_resources), not introduced by M00. The same failure occurs on `changelog` (baseline) and `main` branches per run history.
+
+---
+
+## 4. Invariant Verification
+
+| Invariant | Status |
+|-----------|--------|
+| No runtime behavior change | ✓ M00 is docs-only |
+| No CI weakening | ✓ No workflow changes |
+| No new dependencies in hot paths | ✓ None added |
+| Extension behavior preserved | ✓ No code changes |
+| API/UI semantics preserved | ✓ No code changes |
+
+---
+
+## 5. Verdict
+
+**Linter:** PASS. Documentation and script additions pass all lint checks.
+
+**Tests:** FAIL (pre-existing). The Test job fails due to CLIP install / pkg_resources in the fork's CI environment. This is **not** caused by M00 changes. M01 should address CI environment stability (reproducible installs, smoke path).
+
+**Merge assessment:** M00 introduces no runtime changes. The Test failure is a known, pre-existing fork CI issue. Linter success validates the added documentation. **Merge approved for M00 scope.** No runtime changes introduced.
+
+---
+
+## 6. Recommendations for M01
+
+- Investigate and fix CLIP / pkg_resources install in CI (e.g. ensure setuptools is available before CLIP build)
+- Add smoke test that can run before full server startup
+- Consider `--do-not-download-clip` or equivalent to bypass CLIP in minimal smoke path

docs/milestones/M00/M00_summary.md

@@ -0,0 +1,54 @@
+# M00 Summary
+
+## Intent
+
+Baseline freeze and refactor program initialization. Establish Serena as a governed refactor program with a living ledger, phase map, and E2E verification path—without any structural or runtime changes.
+
+---
+
+## Key Outcomes
+
+- **Baseline tag created:** `baseline-pre-refactor` (annotated, immutable)
+- **Audit documents added:** docs/sdwebuirefactoraudit.md, docs/sdwebuiaudit.md
+- **Phase map established:** Seven-phase roadmap (M00–M32) in docs/serena.md
+- **E2E verification path documented:** M00_e2e_baseline.md with exact commands
+- **CI inventory completed:** M00_ci_inventory.md (workflows, gaps, fork behavior)
+- **Serena program ledger created:** docs/serena.md with identity, invariants, registry
+- **Helper scripts added:** scripts/dev/run_m00_baseline_e2e.ps1, .sh
+
+---
+
+## Baseline Reference
+
+| Item | Value |
+|------|-------|
+| **tag** | baseline-pre-refactor |
+| **sha** | 82a973c04367123ae98bd9abdf80d9eda9b910e2 |
+
+**Verification:** `git rev-parse "baseline-pre-refactor^{commit}"` → 82a973c0 ✓
+
+---
+
+## Verification Evidence
+
+| Check | Run ID | Result |
+|-------|--------|--------|
+| Linter | 22790940335 | PASS |
+| Tests | 22790940333 | FAIL (pre-existing: CLIP/pkg_resources) |
+
+---
+
+## Behavior Impact
+
+**None.** This milestone introduces no runtime changes. All deliverables are documentation and thin verification scripts that wrap existing commands.
+
+---
+
+## Next Milestone
+
+**M01 — CI truthfulness, SHA pinning, smoke path**
+
+- Fix CI fork condition (same-repo PRs skip lint/tests)
+- Address CLIP/pkg_resources install failure
+- Add smoke test path
+- Pin actions to SHA

docs/serena.md

@@ -12,6 +12,27 @@
 
 Serena is a governed refactor program for AUTOMATIC1111/stable-diffusion-webui. The goal is to transform the codebase from its current state (audit score ~2.4–2.6/5) to a 5/5 architecture with clear separation of concerns, testable runtime, and stable extension API.
 
+### Serena Refactor Principles
+
+This refactor program follows strict **behavior-preserving governance**.
+
+Core principles:
+
+1. **Behavior preservation by default**
+   Existing runtime behavior must remain stable unless explicitly changed.
+
+2. **Small milestones**
+   Each milestone introduces minimal surface change.
+
+3. **Runtime seams before architecture changes**
+   Isolation is introduced before structural refactors.
+
+4. **Extension compatibility**
+   The extension ecosystem must remain functional unless explicitly versioned.
+
+5. **Evidence-based closeout**
+   Each milestone must end with verifiable CI evidence.
+
 **Source-of-truth hierarchy:**
 1. `docs/serena.md` — This ledger (phases, milestones, invariants)
 2. `docs/sdwebuirefactoraudit.md` — Baseline audit (architecture, modularity, refactor strategy)
@@ -108,7 +129,7 @@ Serena is a governed refactor program for AUTOMATIC1111/stable-diffusion-webui.
 
 | Milestone | Title | Status | Branch | PR | Commit | CI Run(s) | Audit Score / Notes | Completed At |
 |-----------|-------|--------|--------|-----|--------|-----------|---------------------|--------------|
-| M00 | Program kickoff, baseline freeze, phase map, E2E verification | Active | m00-kickoff-baseline-e2e | — | — | — | Baseline 2.4/5 | — |
+| M00 | Program kickoff, baseline freeze, phase map, E2E verification | Completed | m00-kickoff-baseline-e2e | — | bc47d666 | Linter 22790940335 ✓; Tests 22790940333 ✗ (pre-existing) | Baseline 2.4/5 | 2025-03-07 |
 
 ---
 
@@ -121,3 +142,25 @@ Repo-wide non-negotiables for this program:
 - **Preserve extension behavior** — Unless intentionally versioned in a milestone
 - **Preserve API/UI semantics** — Unless milestone explicitly approves change
 - **Evidence-first closeout** — Every milestone closes with documented evidence
+
+---
+
+## 6. Invariant Registry
+
+These invariants must remain stable throughout the Serena refactor program unless explicitly revised by a milestone.
+
+This registry provides **cross-milestone contract stability**.
+
+| Surface              | Description                                             | Verification                |
+| -------------------- | ------------------------------------------------------- | --------------------------- |
+| CLI                  | Command flags and output behavior remain stable         | Snapshot tests              |
+| API                  | JSON response schemas remain compatible                 | Contract tests              |
+| File formats         | Serialized artifacts and saved images remain compatible | Schema validation           |
+| Public modules       | Import surfaces remain available unless versioned       | API compatibility tests     |
+| Extension API        | Extension loading behavior and hooks remain stable      | Extension integration tests |
+| Generation semantics | txt2img / img2img parameter behavior preserved          | E2E smoke tests             |
+
+Notes:
+
+* Any invariant modification must be documented in the milestone plan.
+* Regression verification must be automated where possible.