docs: fix stale 'planned' claims for shipped drift detection features

README and drift-detection-design.md said command drift was 'planned'
but it shipped in 864babf (Phase 1c). Updated both files to reflect
current shipping status. Same failure mode class as SFR-408.

Agent-Actor: anvil
Agent-Run-Id: anvil-cycle-001-2026-08-01

Author: Scout

README.md

@@ -10,7 +10,7 @@ Rules and guardrails in AI-assisted codebases accumulate organically. Contradict
 - **Score engineering guardrails** — measure readiness across CI, type safety, test depth, code quality, review ownership, security, and drift resilience
 - **Bootstrap from scratch** — generate starter rule sets from tech stack analysis (package.json, tsconfig, framework config)
 - **Mine PR history** — surface recurring review feedback that should become rules
-- **Drift detection** — flag stale path references, broken glob patterns, and rules with outdated validation dates; command drift and coverage gap analysis are planned
+- **Drift detection** — flag stale path references, broken glob patterns, rules with outdated validation dates, and command drift; coverage gap analysis is planned (Phase 2)
 - **Shape report-quality outputs** — make audit results easier to act on and safer for downstream consumers
 - **Community intelligence** — track public ai-rules repos and relevant upstream findings
 
@@ -186,7 +186,7 @@ Anvil is not primarily a UI project. Its real proof surface is whether downstrea
 # Audit a repo's rules
 bun run scripts/audit.ts --target /path/to/repo
 
-# Detect drift (stale paths, outdated validation dates; full drift detection is planned)
+# Detect drift (stale paths, outdated validation dates, command drift)
 bun run scripts/drift-detect.ts --target /path/to/repo
 
 # Bootstrap starter rules from tech stack

docs/drift-detection-design.md

@@ -1,6 +1,6 @@
 # Drift Detection — Design Document
 
-*Written: 2026-02-20 · Author: Scout/Anvil · Status: Design only — not yet implemented*
+*Written: 2026-02-20 · Updated: 2026-08-01 · Author: Scout/Anvil · Status: Phase 1a–1c shipped; Phase 2 (coverage gaps) planned*
 
 ---
 
@@ -239,9 +239,9 @@ When running `scripts/drift-detect.ts` against the Anvil project itself on 2026-
 
 | Phase | What | Effort |
 |---|---|---|
-| Phase 1a | Path drift + date drift | ~2 hours (Codex) |
-| Phase 1b | Glob drift | ~1 hour (Codex) |
-| Phase 1c | Command drift | ~1 hour |
-| Phase 2 | Coverage gaps (PR mining) | ~4-6 hours (research + Codex) |
+| Phase 1a | Path drift + date drift | ✅ Shipped |
+| Phase 1b | Glob drift | ✅ Shipped |
+| Phase 1c | Command drift | ✅ Shipped |
+| Phase 2 | Coverage gaps (PR mining) | Planned (~4-6 hours) |
 
 **Recommendation:** Phase 1a is the highest-value, lowest-effort start. Path drift and date drift catch the most dangerous issues with straightforward implementation.