2 Commits

Author	SHA1	Message	Date
Garry Tan	b515f31400	feat(security): always run Haiku on tool outputs (drop the L4 gate) ... Tool-result scan previously short-circuited when L4 (TestSavantAI) scored below WARN, and further gated Haiku on any layer firing at >= LOG_ONLY. On BrowseSafe-Bench that meant Haiku almost never ran, because TestSavantAI has ~15% recall on browser-agent-specific attacks (social engineering, indirect injection). We were gating our best signal on our weakest. Run all three classifiers (L4 + L4c + Haiku) in parallel. Cost: ~$0.002 + ~8s Haiku wall time per tool result, bounded by the 15s Haiku timeout. Haiku also runs in parallel with the content scans so it's additive only against the stream handler budget, not against the session wall time. User-input pre-spawn path unchanged — shouldRunTranscriptCheck still gates there. The Stack Overflow FP mitigation that original gate was built for still applies to direct user input; tool outputs have different characteristics. Source-contract test updated to pin the new parallel-three shape. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-04-20 21:15:57 +08:00
Garry Tan	9bbfa26597	test(security): source-level contracts for the security wiring ... 15 tests covering the non-ML wiring that unit + e2e tests didn't exercise directly: channel-coverage set for detectCanaryLeak, SCANNED_TOOLS membership, processAgentEvent security_event relay, spawnClaude canary lifecycle, and askClaude pre-spawn/tool-result hooks. Generated by /ship coverage audit — 87% weighted coverage. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-04-20 07:09:52 +08:00