ypollak2/llm-router

Ruff Linting Cleanup

Fixed 14 ruff linting violations from refactoring fallout:

• Removed unused imports across hook_client, service, service_manager, and test files
• Eliminated unused variable assignments
• Code now passes uvx ruff check src/ tests/ with 0 errors

New: Comprehensive v5.3 Audit Demo

Created tests/test_demo_v53.py with 16 tests validating core v5.3 features:

Demo 3a — Heuristic Classifier (5 tests)

Fast local classification without HTTP, covering infrastructure, query, code, analysis, and generation detection

Demo 3b — Sidecar Service HTTP API (5 tests)

FastAPI service endpoints validation: /health readiness and /classify routing decisions

Demo 3c — Enforce-Route Observation Mode (3 tests)

Validates enforce-route hook documentation and core tool protection design

Demo 3d — Correlation ID Propagation (3 tests)

Tests correlation ID generation and response structure integration

Release Scripts

Added scripts/audit_demo.sh — automated verification script that:

• Runs ruff linting checks
• Executes the full demo test suite
• Runs critical path regression tests
• Outputs release-ready summary

Quality Metrics

• All 913 tests passing (including 16 new demo tests)
• Ruff fully clean across src/ and tests/
• Full backward compatibility maintained

Upgrade

```bash
pip install --upgrade claude-code-llm-router && llm-router install
```