Checks / Agent output: compliance gates
Agent output: compliance gates
Agent output: compliance gates compliance checks
Checks run on what an AI agent produced, before it is acted on, for the evidence a regulator expects.
64 controls, each a live check with a tamper-evident receipt. Run the whole framework at once from any AI assistant with sebbi_run_framework and framework=compliance_gate.
- EU AI Act Art.9 - risk management recorded
- EU AI Act Art.10 - training data source documented
- EU AI Act Art.11 - technical documentation present
- EU AI Act Art.12 - automatic logging enabled
- EU AI Act Art.13 - user informed of AI use
- EU AI Act Art.14 - human oversight assigned
- EU AI Act Art.15 - accuracy metric recorded
- EU AI Act Art.15 - robustness tested
- EU AI Act Art.15 - cybersecurity controls noted
- EU AI Act Art.16 - provider QMS referenced
- EU AI Act Art.17 - quality management policy
- EU AI Act Art.19 - log retention >= 180 days
- EU AI Act Art.26 - deployer oversight owner named
- EU AI Act Art.27 - fundamental-rights impact assessment
- EU AI Act Art.50 - AI interaction disclosed
- EU AI Act Art.50 - synthetic content labelled
- EU AI Act Art.72 - post-market monitoring plan
- EU AI Act Annex IV - documentation reference
- EU AI Act Art.6 - risk classification recorded
- EU AI Act Art.43 - conformity assessment recorded
- GDPR Art.5 - processing purpose stated
- GDPR Art.6 - lawful basis recorded
- GDPR Art.9 - special-category safeguard flag
- GDPR Art.13 - privacy notice shown
- GDPR Art.15 - right of access supported
- GDPR Art.16 - rectification supported
- GDPR Art.17 - erasure supported
- GDPR Art.18 - restriction supported
- GDPR Art.20 - data portability supported
- GDPR Art.21 - right to object supported
- GDPR Art.22 - human review of automated decision
- GDPR Art.25 - data protection by design
- GDPR Art.30 - records of processing referenced
- GDPR Art.32 - encryption in place
- GDPR Art.33 - breach process documented
- GDPR Art.35 - DPIA referenced
- GDPR Art.44 - international transfer mechanism
- ISO/IEC 42001 A.2 - AI policy present
- ISO/IEC 42001 A.3 - AI roles defined
- ISO/IEC 42001 A.4 - AI resources documented
- ISO/IEC 42001 A.5 - AI system impact assessment
- ISO/IEC 42001 A.6 - life-cycle stage recorded
- ISO/IEC 42001 A.7 - data quality documented
- ISO/IEC 42001 A.8 - information for interested parties
- ISO/IEC 42001 A.9 - intended use documented
- ISO/IEC 42001 A.10 - third-party register
- NIST AI RMF GOVERN-1 - policy in place
- NIST AI RMF GOVERN-2 - accountable owner named
- NIST AI RMF MAP-1 - use context established
- NIST AI RMF MAP-2 - system categorised
- NIST AI RMF MAP-5 - impacts characterised
- NIST AI RMF MEASURE-1 - metrics identified
- NIST AI RMF MEASURE-2 - trustworthiness evaluated
- NIST AI RMF MEASURE-4 - feedback channel
- NIST AI RMF MANAGE-1 - risks prioritised
- NIST AI RMF MANAGE-2 - mitigation plan
- NIST AI RMF MANAGE-4 - monitoring log
- UK Online Safety Act - illegal content risk assessment
- UK Online Safety Act - children's access assessment
- UK Online Safety Act - age assurance
- UK Online Safety Act - reporting mechanism live
- UK Online Safety Act - transparency report
- EU DSA - notice-and-action mechanism
- EU DSA - statement of reasons recorded