4 months ago
9c078c9- modelRouter: 42 tests over routing branches (fixed-model, cross-provider allowlist/connectivity, LM Studio fallback, reasoning-effort precedence) - toolPolicy: 22 tests for role/profile tool filtering incl. MCP gating - roles: 23 tests for role definitions, shell policy, prompt lines - reportParser: empty/multiple/legacy/garbage <agent_report> cases - sync queue: 55 tests (JSONL round-trip, caps, backoff, dedupe) - sync redaction: 57 tests (allowlist, secret/path patterns, payload sanitize) - repl command router: 35 tests across 13 slash commands Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Parent442a1c8