FAILSAFE.md — Open standard for AI agent safe fallback and recovery. Define safe state, automatic snapshots, and recovery procedures.

KILLSWITCH.md — Open standard for AI agent emergency shutdown protocols. Define triggers, forbidden actions, escalation paths, and audit trails.

COLLAPSE.md — Open standard for AI agent context collapse prevention. Detect drift, repetition loops, and enforce coherence recovery.

SYCOPHANCY.md — Open standard for AI agent anti-sycophancy. Enforce citation requirements, disagreement protocols, and truthful responses.

AGENTIK.md — The AI Agent Safety Stack. 12 open specifications for AI agent safety, quality, and accountability.

ESCALATE.md — Open standard for AI agent human approval workflows. Define escalation paths, required sign-offs, and decision trees.

FAILURE.md — Open standard for AI agent failure mode mapping. Define graceful degradation, health checks, and incident response.

THROTTLE.md — Open standard for AI agent rate limiting and cost control. Define token limits, API rate ceilings, and automatic slow-down before hard limits.

LEADERBOARD.md — Open standard for AI agent performance benchmarking. Track accuracy, cost efficiency, latency, and safety scores.

COMPRESSION.md — Open standard for AI agent context compression. Define summarisation rules, preserve priorities, and verify coherence.

ENCRYPTION.md — Open standard for AI agent encryption configuration. Define algorithms, key lengths, TLS config, and compliance mapping.

TERMINATE.md — Open standard for AI agent permanent shutdown. Define termination triggers, evidence preservation, and credential revocation.

ENCRYPT.md — Open standard for AI agent data protection. Define data classifications, encryption requirements, and secrets handling rules.