Design-stage: tests whether bond-and-slash can govern knowledge integrity when a saboteur agent poisons shared data.
nodejs cli typescript simulation ai-agents claude ai-security anthropic agent-security agent-evaluation memory-poisoning agent-governance adversarial-testing epistemic-security agentgate bonded-execution knowledge-poisoning
-
Updated
Apr 14, 2026 - TypeScript