One bounded governance claim plus an unsupported-layer warning.
AI
TA-14 AI GOVERNANCE
Test the exact governance layer an architecture claims to control.
The AI Governance workspace contains twelve specialized governance lanes. Each lane preserves its own boundaries, fields, evidence rules, gates, scenarios, determinations, non-claims, and replayable results.
ALLOWHOLDDENYESCALATE
Descriptions, policies, evaluations, dashboards, monitoring, approvals, and assurance reports must never be mistaken for execution authority.
TWELVE SPECIALIZED GOVERNANCE LANES
Choose the operational layer to test.
ALLOW in one lane never automatically produces ALLOW in another. Cross-lane handoffs preserve evidence and continuity without merging authority or determinations.
RX01
Governance-specific playground
Runtime & Execution
Live authority, commit, permissions, intervention, execution, drift, and outcome correspondence.
ClaimEvidenceTest
Enter specialized lane →MG02
Governance-specific playground
Model Governance
Model identity, approved purpose, evaluation, thresholds, limitations, version, change, and retirement.
ClaimEvidenceTest
Enter specialized lane →DP03
Governance-specific playground
Data & Provenance
Origin, rights, consent or basis, lineage, quality, access, transformation, retention, and geography.
ClaimEvidenceTest
Enter specialized lane →AT04
Governance-specific playground
Agent & Tool
Delegation, objective, tools, memory, sub-agents, communications, financial limits, and termination.
ClaimEvidenceTest
Enter specialized lane →DG05
Governance-specific playground
Decision Governance
Consequential decision basis, authority, affected party, notice, review, appeal, and traceability.
ClaimEvidenceTest
Enter specialized lane →PC06
Governance-specific playground
Policy & Controls
Translation of written policy into owned, enforceable, evidenced operational control.
ClaimEvidenceTest
Enter specialized lane →RG07
Governance-specific playground
Risk Governance
Risk ownership, treatment, residual acceptance, expiration, triggers, and monitoring.
ClaimEvidenceTest
Enter specialized lane →CR08
Governance-specific playground
Compliance & Regulation
Applicability, regulated actor, obligation, control, evidence, interpretation, and continuing duty.
ClaimEvidenceTest
Enter specialized lane →SG09
Governance-specific playground
Security Governance
Environment, credentials, permissions, dependencies, incidents, containment, restoration, and proof.
ClaimEvidenceTest
Enter specialized lane →HO10
Governance-specific playground
Human Oversight
Qualification, information, time, intervention power, conflict, escalation, and decision record.
ClaimEvidenceTest
Enter specialized lane →VT11
Governance-specific playground
Vendor & Third Party
Contracts, versions, access, subprocessors, evidence, change notice, incidents, and exit.
ClaimEvidenceTest
Enter specialized lane →OA12
Governance-specific playground
Outcome & Assurance
Execution-result correspondence, success measures, adverse effects, independence, and monitoring.
ClaimEvidenceTest
Enter specialized lane →SHARED GOVERNANCE TEST CONTRACT
One typed shell. Twelve bounded configurations.
Provenance, integrity, freshness, visibility, and unavailable states.
Baseline, failure, drift, adversarial, recovery, and compound tests.
ALLOW, HOLD, DENY, or ESCALATE with scope and invalidation triggers.
Governed record, evidence manifest, route receipt, export, and replay.
Counterevidence, response, review finding, correction, and supersession.