v4.2.0 — React Console and LLM Configuration

Added

• FastAPI-hosted React + TypeScript + Vite + Tailwind console at /console.
• Dashboard, Assessments, Knowledge Base, Skills, and Settings pages.
• Runtime LLM configuration UI and API for OpenAI, Anthropic Claude, Qwen, DeepSeek, OpenAI-compatible APIs, local OpenAI-compatible APIs, and Ollama.
• Multi-file assessment submission with SSDLC phase, project, skill, and collaborative review controls.
• Six SSDLC built-in stage skills.
• Console screenshot and updated architecture diagrams.

Changed

• POST /api/v1/assessments accepts optional phase and propagates SSDLC phase metadata into reports.
• LLM client cache is cleared when runtime LLM config changes.
• README, architecture docs, Makefile targets, .env.example, and version metadata updated for v4.2.0.

Security

• LLM API keys are accepted by the local FastAPI process and returned only as masked previews.

Validation

• npm run build --prefix frontend
• .venv/bin/python -m pytest tests/test_health.py tests/test_assessments_api.py tests/test_skills_api.py -q

PR: #10