Rewrite Readiness with Signal Engine

[digitarald]

"Rewrite Readiness with Signal Engine" — Replace Primer's criteria-based readiness internals with the Azure fork's signal/detector/recommender engine.

9 Phases:

1. Signal Engine Foundation — Core types (Signal, Recommendation, Policy, Detector, Recommender) + executePolicies() engine + PolicyContext implementation
2. Repo Health Detectors — Convert 16 existing criteria into signal-emitting detectors (lint-config, typecheck-config, build-script, ci-config, test-script, etc.)
3. AI Tooling Detectors — Port 5 Azure fork detectors with 40+ patterns across all AI tools (Copilot, Claude Code, Cursor, Windsurf, Cline, Roo Code, Aider, etc.)
4. Recommenders — Port 5 recommenders (missing files, invalid files, settings, location, MCP)
5. Default Policy & Bridge Layer — Wire it together; bridge maps signals → ReadinessCriterionResult[] for backward compat. Detailed signal→criterion mapping table defined.
6. Policy System Adapter — Wrap PolicyConfig as Policy internally via wrapPolicyConfig() adapter
7. Command & Output Updates — --verbose, --include-personal flags, recommendations in console/JSON/HTML
8. VS Code Extension Updates — Recommendations tree section, AI Tooling Score
9. Tests — Engine, detector, recommender, bridge tests + update existing readiness tests

Key Design Rules:

• Signals are observations, not verdicts (detectors never compute pass/fail)
• PolicyContext includes RepoAnalysis (required for repo health detectors)
• ReadinessReport type preserved with additive fields (signals[], recommendations[], score, grade)