P12-T3: Add Error Classification & Categorization

[SoundBlaster]

Description

Implements P12-T3: error classification and categorization for JSON-RPC responses.

• Adds categorize_error(code) helper mapping error codes to buckets: protocol (-32600…-32699), timeout (-32001), tool (≥1), unknown
• Extends MetricsCollector and SharedMetricsStore with error_counts_by_code tracking
• Extends record_response() in both stores to accept error_code and error_message
• Adds get_error_code() / get_error_message() accessors to MCPResponse in schemas.py
• Adds error_code / error_message columns to SQLite requests table with backward-compatible migration
• Passes error_code and error_message through __main__.py response handler and AuditLogger
• Adds error breakdown doughnut chart to dashboard (chart-error-breakdown canvas)
• Color-codes audit table error column by severity: .error-protocol (red), .error-tool (orange), .error-timeout (yellow), .error-unknown (grey)

Follow-up tasks created: FU-P12-T3-1 (document unused error_message in in-memory collector), FU-P12-T3-2 (add error_code to CSV audit export).

Type of Change

• Bug fix
• New feature
• Documentation update
• Refactoring
• CI/CD improvement

Quality Gates

Before submitting, ensure all quality gates pass:

make check

Or run individually:

• make test - All tests pass with ≥90% coverage — 437 passed, 5 skipped; 96.09% coverage
• make lint - No linting errors — ruff check src/: all checks passed
• make format - Code is properly formatted
• make typecheck - Type checking passes — mypy src/: no issues in 13 source files
• make doccheck - Documentation is synced with DocC (if docs changed)

Documentation Sync

• Documentation changes are synced with DocC catalog (or N/A) — no docs/ changes

Testing

• Added/updated tests for new functionality
  • TestCategorizeError (9 tests) — boundary coverage for all categories
  • TestParseErrorInfo (5 tests) — success, error with code/message, invalid JSON
  • test_error_counts_by_code_* (5 tests in test_metrics.py, 5 in test_shared_metrics.py)
• All tests pass locally
• Manually tested the changes

Checklist

• Code follows the project's style guidelines
• Self-review completed
• Comments added for complex code
• Documentation updated (if needed)
• No new warnings generated
• PR title is descriptive