feat: implement v0.7.0 AI agent enhancements

Implements comprehensive AI agent improvements with enhanced reporting,
CLI enhancements, quality controls, and testing.

## Phase 1: Enhanced Reporting
- Add AI findings section to HTML reports with perspective tabs
- Add AI findings section to Markdown reports with metadata
- Restore HTML template with AI findings visualization
- Update HTML report generator to extract and categorize AI findings

## Phase 2: CLI Enhancements
- Add simctl ai test command for backend validation
- Add simctl ai list command for agent enumeration
- Fix simctl agents list --type=ai with config support
- Enhanced agent listing with detailed metadata

## Phase 3: Quality Controls
- Implement finding deduplication between deterministic and AI agents
- Add confidence filtering for AI findings (configurable threshold)
- Integrate deduplication and filtering into orchestrator engine
- Add comprehensive unit tests for both features

## Phase 4: Documentation
- Update quickstart guide with AI agent examples
- Add backend comparison (MLX vs llama.cpp vs remote vs API)
- Add troubleshooting guide for common issues

## Phase 5: Testing & Validation
- Add unit tests for deduplication and filtering
- Create end-to-end testing script for real models
- Verify CLI commands and functionality
- Fix all linter errors and compilation issues

## Technical Improvements
- Add ExistingFindings field to rpc.Task for AI agent context
- Fix unused parameter warnings in engine.go
- Remove duplicate definitions in cmd/simctl
- Add comprehensive test coverage

Implements: v0.7.0 AI Agent Enhancements
Status: Complete ✅