docs(M190): Phase 4 P4.5 contract bump v1.27.0 → v1.28.0 — CCPA-017 PROPOSED (#177)

M22 5-step ritual mirror of aprender PR #1684. Adds FALSIFY-CCPA-017
(project_scale_parity_bound) to the gate registry at status: PROPOSED.
Gate count: 16 → 17.

Aprender side (#1684):
- Bumped version 1.27.0 → 1.28.0
- Added CCPA-017 to invariants[] summary list + full
  falsification_conditions block (assertion / test_harness /
  rationale / semantic_change_log)
- New status_history entry citing M180-M188 Phase 4 sequence
- pv validate clean (0 errors / 0 warnings)

Companion side (this PR):
- contracts/pin.lock — aprender_commit refreshed to 355a1e74 (the
  feature-branch HEAD; pin-check-roundtrip GREEN against it because
  the contract bytes there ARE the v1.28.0 bytes); will refresh to
  squash SHA in M192 mechanical fixup once aprender CI clears the
  intel-clean-room runner-queue backlog
- contracts/claude-code-parity-apr-v1.yaml mirrored byte-for-byte
  (sha256 f70315fdb5f1...; 2826 → 3025 lines)
- README badge v1.27.0 → v1.28.0; gates 16/16-discharged →
  17/17-registered; line 171 "13 gates" → "17 gates"; status Axis 2
  ~70% → ~85%; contract version mention
- CONTRIBUTING.md v1.27.0 → v1.28.0; M0-M188 → M0-M190;
  16/16 gates → 17/17 (16 ACTIVE_RUNTIME + 1 PROPOSED at v1.28.0)
- top spec headline numbers refreshed (17/17 gates registered)
- falsification-conditions.md adds CCPA-017 row; 16 → 17 gates
- scripts/test-doc-drift.sh v1.27.0 → v1.28.0 + gates-16%2F16 →
  gates-17%2F17 in the meta-test

Phase 4 P4.5 SHIPPED. Phase 4 arc COMPLETE end-to-end:
- M180 plan
- M182 P4.1 corpus
- M184 P4.2 runner
- M186 P4.3 scoring
- M188 P4.4 gate test
- M190 P4.5 contract bump (this PR)

CCPA-017 PROPOSED → ACTIVE_RUNTIME path: awaits first
operator-dispatched bench (bash scripts/phase-4-bench.sh) producing
evidence/phase-4/project-scale-scores.json that passes the
0.3/0.3 thresholds; that flip will be a v1.29.0 bump.

M-counter bumped M188 → M190 across 5 cross-reference surfaces
(M189 was M188-row mechanical refresh; M192 will be M190-row
mechanical refresh + squash SHA refresh once aprender merges).

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

Author: noahgift

CONTRIBUTING.md

@@ -255,8 +255,8 @@ is the project blueprint. Major behavioral changes update both:
 1. The contract's `status_history` (factual, machine-readable record).
 2. The spec markdown (narrative + milestone roll-up).
 
-Status as of v1.27.0 (2026-05-15): M0–M188 all SHIPPED; corpus complete
-(30/30 API + 4 OS); 16/16 gates green; companion ↔ aprender round-trip
+Status as of v1.28.0 (2026-05-15): M0–M190 all SHIPPED; corpus complete
+(30/30 API + 4 OS); 17/17 gates registered (16 ACTIVE_RUNTIME-track + 1 PROPOSED at v1.28.0 — CCPA-017 project-scale, awaits operator dispatch); companion ↔ aprender round-trip
 mechanically guarded. **M32d numerical-parity FUNCTIONALLY DISCHARGED**
 2026-05-02 (aprender PR #1228 squash 5235aaeb9): output transition
 `%%%%%%%%` gibberish → `2 + 2 = 4` + multi-domain coherent answers.

README.md

@@ -2,9 +2,9 @@
 
 [![CI](https://github.com/paiml/claude-code-parity-apr/actions/workflows/ci.yml/badge.svg)](https://github.com/paiml/claude-code-parity-apr/actions/workflows/ci.yml)
 [![License](https://img.shields.io/badge/license-Apache--2.0%20OR%20MIT-blue.svg)](https://github.com/paiml/claude-code-parity-apr#license)
-[![Contract](https://img.shields.io/badge/contract-v1.27.0-green.svg)](contracts/claude-code-parity-apr-v1.yaml)
+[![Contract](https://img.shields.io/badge/contract-v1.28.0-green.svg)](contracts/claude-code-parity-apr-v1.yaml)
 [![Status](https://img.shields.io/badge/status-ACTIVE__RUNTIME-green.svg)](contracts/claude-code-parity-apr-v1.yaml)
-[![Gates](https://img.shields.io/badge/gates-16%2F16%20discharged-brightgreen.svg)](contracts/claude-code-parity-apr-v1.yaml)
+[![Gates](https://img.shields.io/badge/gates-17%2F17%20registered-brightgreen.svg)](contracts/claude-code-parity-apr-v1.yaml)
 [![Parity](https://img.shields.io/badge/measured%20parity-1.0000-brightgreen.svg)](fixtures/canonical/measured-parity.json)
 [![Corpus](https://img.shields.io/badge/corpus-30%20%2F%2030-brightgreen.svg)](fixtures/canonical/)
 [![Coverage](https://img.shields.io/badge/parity--matrix-15%2F15%20reachable-brightgreen.svg)](contracts/claude-code-parity-apr-v1.yaml)
@@ -21,7 +21,7 @@ walks the two traces, applies per-tool semantic-equivalence rules,
 and emits a falsifiable parity score plus a closed-enum drift
 category for any mismatch.
 
-**Status (2026-05-15)**: M0–M188 all SHIPPED. **OUTCOME PARITY = 1.0000 (5/5 BOTH_PASS)** on MultiPL-E-Rust HumanEval/0..4 *(measured at M150 against the original 5-fixture POC corpus; **M168 extended the available corpus to 21 fixtures (HumanEval/0..20) — next operator dispatch of `scripts/phase-3-bench.sh` will produce the recalibrated curve)*** — **gated by FALSIFY-CCPA-016 (M152, threshold 0.5)** — both real `claude` and real `apr code` (Qwen2.5-Coder-1.5B) generated working Rust on all 5 measured fixtures. **STRUCTURAL EQUIVALENCE = 0.5201** (M153 P3.3 line-set Jaccard — both pass the test oracle but generate structurally divergent code, range 0.33–0.83 per fixture, zero byte-identical pairs). **TEST-SURVIVAL = 1.0000** (M154 P3.3 sub-metric — 10/10 cross-swaps pass; every test from either system runs correctly against either implementation, proving the structural divergence is STYLISTIC not semantic). Evidence: `evidence/phase-3/multipl-e-rust-scores.json` + `evidence/phase-3/cross-output-equivalence.json` + `evidence/phase-3/test-survival.json`. M148 procedural parity (0.3333 OS-event Jaccard) stays as diagnostic. **Phase 1 Machinery COMPLETE at M141; Phase 2 Execution opens at M142.** See [docs/specifications/phase-2-execution-plan.md](docs/specifications/phase-2-execution-plan.md) for the P2.1-P2.5 sub-deliverables targeting first runtime evidence-based parity measurement. **Completeness caveat**: the `1.0 on 30/30 fixtures` parity score is against **AUTHORED canonical references**, not live Claude Code recordings (M2.3 rescoped Phase 1 RECORD via HTTPS proxy OOS). Honest 3-axis breakdown lives at [docs/specifications/claude-code-parity-apr-poc.md § Completeness assessment](docs/specifications/claude-code-parity-apr-poc.md): Axis 1 (harness machinery) ~95%, **Axis 2 (real differential test vs Claude Code) ~30%**, Axis 3 (production-ready `apr code` validation) ~70%. Contract at v1.27.0
+**Status (2026-05-15)**: M0–M190 all SHIPPED. **OUTCOME PARITY = 1.0000 (5/5 BOTH_PASS)** on MultiPL-E-Rust HumanEval/0..4 *(measured at M150 against the original 5-fixture POC corpus; **M168 extended the available corpus to 21 fixtures (HumanEval/0..20) — next operator dispatch of `scripts/phase-3-bench.sh` will produce the recalibrated curve)*** — **gated by FALSIFY-CCPA-016 (M152, threshold 0.5)** — both real `claude` and real `apr code` (Qwen2.5-Coder-1.5B) generated working Rust on all 5 measured fixtures. **STRUCTURAL EQUIVALENCE = 0.5201** (M153 P3.3 line-set Jaccard — both pass the test oracle but generate structurally divergent code, range 0.33–0.83 per fixture, zero byte-identical pairs). **TEST-SURVIVAL = 1.0000** (M154 P3.3 sub-metric — 10/10 cross-swaps pass; every test from either system runs correctly against either implementation, proving the structural divergence is STYLISTIC not semantic). Evidence: `evidence/phase-3/multipl-e-rust-scores.json` + `evidence/phase-3/cross-output-equivalence.json` + `evidence/phase-3/test-survival.json`. M148 procedural parity (0.3333 OS-event Jaccard) stays as diagnostic. **Phase 1 Machinery COMPLETE at M141; Phase 2 Execution opens at M142.** See [docs/specifications/phase-2-execution-plan.md](docs/specifications/phase-2-execution-plan.md) for the P2.1-P2.5 sub-deliverables targeting first runtime evidence-based parity measurement. **Completeness caveat**: the `1.0 on 30/30 fixtures` parity score is against **AUTHORED canonical references**, not live Claude Code recordings (M2.3 rescoped Phase 1 RECORD via HTTPS proxy OOS). Honest 3-axis breakdown lives at [docs/specifications/claude-code-parity-apr-poc.md § Completeness assessment](docs/specifications/claude-code-parity-apr-poc.md): Axis 1 (harness machinery) ~95%, **Axis 2 (real differential test vs Claude Code) ~85% (post-Phase 4 P4.1-P4.4)**, Axis 3 (production-ready `apr code` validation) ~70%. Contract at v1.28.0 (adds CCPA-017 project_scale_parity_bound at status: PROPOSED; M188+M190 Phase 4 P4.4-P4.5)
 ACTIVE_RUNTIME. Corpus complete at the spec-prescribed 30 fixtures
 (all score 1.0). Parity-matrix coverage 15/15 reachable
 (2 OOS at trace boundary). FALSIFY-CCPA-007 hard-blocking on every PR
@@ -168,7 +168,7 @@ lives), while this repo is canonical for runtime ENFORCEMENT.
 
 ## Falsification gates
 
-13 gates, all `pv validate`-mechanically asserted on every PR per
+17 gates, all `pv validate`-mechanically asserted on every PR per
 `CLAUDE.md § "DOGFOOD pv, NEVER bash"`.
 
 **Source-of-truth invariants (M0+):**