Agentic Covenants / Agentic Covenants Matrix
Approval gating at In-agent
This layer is advisory.
Model says "are you sure?" Silently bypassable. Anthropic Auto Mode (March 26, 2026): 93% approval rate on permission prompts is the empirical ceiling for human-in-the-loop on top of in-agent gates. Treat any approval prompt that fires often as already broken.
How it is still bypassed
- the model can be talked out of asking
Maps to
- NIST CSF 2.0
- (advisory)
- NIST AI RMF
- MANAGE 4.1
- OWASP LLM
- LLM06
- OWASP Agentic
- ASI09