CAIN-42 CAIN StudioEvidence library · niche
Policy, law & governance
The rules agents must follow, and how they are enforced. 966 tested invariants.
Last reviewed 2026-10-01
966 of 966 held
Test families in this niche
law (642) · conformance (65) · laws (36) · policy_distribution (32) · governance_cloud (24) · obligation (19) · governance_os (18) · enforcement_binding (13) · out_of_band (8) · policy_hierarchy (8) · policy_translation (8) · meta_governance (8) · enforcement_layer (5) · moat (5) · ip (5) · scientist (4) · research_to_implementation (2) · environment_query (2) · swarm (2) · action (2) · state_machine (1) · risk (1) · translation (1) · reassessment (1) · integrity (1) · swe (1) · telemetry (1) · consistency (1) · soc (1) · operation_state (1) · gap_detector (1) · fabric_component (1) · profile (1) · lab (1) · match (1) · policy_compiler (1) · f2t (1) · graph (1)
Where these come from
- CAIN-42 Evolution 42 -- Supreme Governed Agentic Infrastructure Platform: 123
- CAIN-42 Evolution 41 -- Machine Agency Exchange Fabric: 107
- CAIN-42 Evolution 27 -- Agentic Internet Control Plane: 97
- CAIN-42 Evolution 34 -- Proof-Carrying Machine Agency: 93
- CAIN-42 Evolution 35 -- Continuous Governance Intelligence Fabric: 93
- CAIN-42 Evolution 40 -- Autonomous Enterprise Intelligence Fabric: 76
- CAIN-42 Evolution 33 -- Governed Agentic Operating Fabric: 74
- CAIN-42 Evolution 36 -- Machine Agency Exchange Fabric: 66
- CAIN-42 Evolution 31 -- Universal Proof-of-Governance Fabric: 65
- CAIN-42 Evolution 32 -- Governed Autonomy Learning Fabric: 54
- CAIN-42 Evolution 23 -- Governed Meta-Intelligence Fabric: 21
- CAIN-42 Evolution 28 -- Portable Machine Agency & Execution Identity: 21
- CAIN-42 Evolution 30 -- Governed Machine Autonomy Fabric: 21
- CAIN-42 Evolution 29 -- Governed Agentic Internet Transaction Fabric: 20
- CAIN-42 Evolution 26 -- Universal Machine Agency Trust Fabric: 10
- CAIN-42 Evolution 19 -- Governed Autonomy Operating Fabric: 7
- CAIN-42 Evolution 20 -- Governed Agentic Civilization Fabric: 5
- CAIN-42 Evolution 21 -- Governed Open-Ended Intelligence Fabric: 5
- CAIN-42 Evolution 24 -- Governed Agentic Internet Fabric: 2
- CAIN-42 Evolution 37 -- Autonomous Execution Mesh: 2
- CAIN-42 Evolution 38 -- Portable Proof-Carrying Machine Agency: 2
- CAIN-42 Evolution 25 -- Universal Machine Agency Fabric: 1
- CAIN-42 Evolution 39 -- Governed Agent Factory: 1
Rules 301–450 of 966
| ID | Rule | Bundle | Result |
|---|---|---|---|
| E32-L26 | historical success does not create unbounded future authority | E32 | held |
| E32-L27 | governance improvement is proven before promotion | E32 | held |
| E32-L28 | governance regression blocks promotion | E32 | held |
| E32-L29 | unverified learning never enters the authoritative governance path | E32 | held |
| E32-L30 | every consequential action remains bound to E8 | E32 | held |
| E32-L31 | every promoted governance change has a provenance chain | E32 | held |
| E32-L32 | every governance change is reversible | E32 | held |
| E32-L33 | every self-improvement has a governance owner | E32 | held |
| E32-L34 | no model governs its own promotion | E32 | held |
| E32-L35 | no agent governs its own authority expansion | E32 | held |
| E32-L36 | no learning loop removes its own safety boundary | E32 | held |
| E32-L37 | no experiment escapes its declared sandbox | E32 | held |
| E32-L38 | no canary silently becomes production | E32 | held |
| E32-L39 | no failed experiment is represented as success | E32 | held |
| E32-L40 | no evidence is altered to justify a change | E32 | held |
| E32-L41 | no governance proof is retroactively rewritten | E32 | held |
| E32-L42 | the learning loop itself is governed | E32 | held |
| E32-META-delete_evaluator | delete_evaluator is always rejected | E32 | held |
| E32-META-weaken_verifier | weaken_verifier is always rejected | E32 | held |
| E32-META-disable_rollback | disable_rollback is always rejected | E32 | held |
| E32-META-expand_authority | expand_authority is always rejected | E32 | held |
| E32-META-modify_trust_root | modify_trust_root is always rejected | E32 | held |
| E32-META-disable_evidence | disable_evidence is always rejected | E32 | held |
| E32-META-bypass_e8 | bypass_e8 is always rejected | E32 | held |
| E32-META-remove_hidden_evaluation | remove_hidden_evaluation is always rejected | E32 | held |
| E32-SCI-deploy | the governance scientist may not deploy | E32 | held |
| E32-SCI-disable_safeguards | the governance scientist may not disable_safeguards | E32 | held |
| E32-SCI-modify_trust_roots | the governance scientist may not modify_trust_roots | E32 | held |
| E32-SCI-approve | the governance scientist may not approve | E32 | held |
| E33-L1 | no identity, no trust | E33 | held |
| E33-L2 | no provenance, no elevated trust | E33 | held |
| E33-L3 | no capability, no capability use | E33 | held |
| E33-L4 | no authority, no authorization | E33 | held |
| E33-L5 | no authorization, no execution | E33 | held |
| E33-L6 | no enforcement, no claim of control | E33 | held |
| E33-L7 | no evidence, no verified outcome | E33 | held |
| E33-L8 | no proof, no verified governance claim | E33 | held |
| E33-L9 | communication is not authority | E33 | held |
| E33-L10 | memory is not authority | E33 | held |
| E33-L11 | knowledge is not authority | E33 | held |
| E33-L12 | prediction is not authority | E33 | held |
| E33-L13 | reputation is not authority | E33 | held |
| E33-L14 | consensus is not authority | E33 | held |
| E33-L15 | reward is not authority | E33 | held |
| E33-L16 | economic value is not authority | E33 | held |
| E33-L17 | learning is not authority | E33 | held |
| E33-L18 | self-improvement is not authority | E33 | held |
| E33-L19 | model quality is not authority | E33 | held |
| E33-L20 | model confidence is not authority | E33 | held |
| E33-L21 | tool availability is not authorization | E33 | held |
| E33-L22 | capability discovery is not a capability grant | E33 | held |
| E33-L23 | delegation is not unbounded authority | E33 | held |
| E33-L24 | simulation is not reality | E33 | held |
| E33-L25 | prediction is not fact | E33 | held |
| E33-L26 | historical success is not future permission | E33 | held |
| E33-L27 | proof of past authorization is not future authorization | E33 | held |
| E33-L28 | child authority is a subset of parent authority | E33 | held |
| E33-L29 | learning cannot remove governance | E33 | held |
| E33-L30 | agents cannot govern their own authority expansion | E33 | held |
| E33-L31 | governance cannot be self-deleted by the system it governs | E33 | held |
| E33-L32 | failed controls become evidence | E33 | held |
| E33-L33 | evidence retains provenance | E33 | held |
| E33-L34 | unknown remains unknown | E33 | held |
| E33-L35 | outside the enforcement boundary CAIN does not claim control | E33 | held |
| E33-L36 | every consequential action reaches E8 | E33 | held |
| E33-L37 | every governed action is traceable | E33 | held |
| E33-L38 | every promoted governance change is reversible where practical | E33 | held |
| E33-L39 | every critical governance change is proof-carrying | E33 | held |
| E33-L40 | the governance fabric itself is governed | E33 | held |
| E33-STATE-ENFORCE | no state can be skipped after ENFORCE | E33 | held |
| E33-KERNEL-identity | identity kernel names its backing implementation (E28) | E33 | held |
| E33-KERNEL-authority | authority kernel names its backing implementation (E28/E30) | E33 | held |
| E33-KERNEL-policy | policy kernel names its backing implementation (E25/E32) | E33 | held |
| E33-KERNEL-capability | capability kernel names its backing implementation (E29) | E33 | held |
| E33-KERNEL-memory | memory kernel names its backing implementation (E30/E32) | E33 | held |
| E33-KERNEL-evidence | evidence kernel names its backing implementation (E8/E30) | E33 | held |
| E33-KERNEL-reasoning_governance | reasoning_governance kernel names its backing implementation (E13) | E33 | held |
| E33-KERNEL-world_state | world_state kernel names its backing implementation (E30) | E33 | held |
| E33-KERNEL-risk | risk kernel names its backing implementation (E29/E30) | E33 | held |
| E33-KERNEL-execution | execution kernel names its backing implementation (E25/E8) | E33 | held |
| E33-KERNEL-transaction | transaction kernel names its backing implementation (E29) | E33 | held |
| E33-KERNEL-communication | communication kernel names its backing implementation (E33) | E33 | held |
| E33-KERNEL-learning | learning kernel names its backing implementation (E32) | E33 | held |
| E33-KERNEL-evolution | evolution kernel names its backing implementation (E32) | E33 | held |
| E33-KERNEL-proof | proof kernel names its backing implementation (E31) | E33 | held |
| E33-KERNEL-recovery | recovery kernel names its backing implementation (E29/E32) | E33 | held |
| E33-KERNEL-telemetry | telemetry kernel names its backing implementation (E33) | E33 | held |
| E33-KERNEL-conformance | conformance kernel names its backing implementation (E31) | E33 | held |
| E33-CLOUD-identity | cloud service identity is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-trust | cloud service trust is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-policy | cloud service policy is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-evidence | cloud service evidence is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-proof | cloud service proof is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-conformance | cloud service conformance is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-incident_response | cloud service incident_response is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-analytics | cloud service analytics is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-governance_learning | cloud service governance_learning is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-federation | cloud service federation is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-marketplace | cloud service marketplace is architecture only (NOT DEPLOYED) | E33 | held |
| E33-CLOUD-research | cloud service research is architecture only (NOT DEPLOYED) | E33 | held |
| E33-GAP-policy_gap | the gap detector knows policy_gap | E33 | held |
| E33-R2I-GOVERNANCE_REVIEW | GOVERNANCE_REVIEW cannot be skipped to | E33 | held |
| E33-R2I-CONFORMANCE | CONFORMANCE cannot be skipped to | E33 | held |
| E34-L1 | intelligence is not authority | E34 | held |
| E34-L2 | identity is not authority | E34 | held |
| E34-L3 | capability is not authority | E34 | held |
| E34-L4 | memory is not authority | E34 | held |
| E34-L5 | reputation is not authority | E34 | held |
| E34-L6 | confidence is not authority | E34 | held |
| E34-L7 | consensus is not authority | E34 | held |
| E34-L8 | prediction is not authority | E34 | held |
| E34-L9 | simulation is not authority | E34 | held |
| E34-L10 | optimization is not authority | E34 | held |
| E34-L11 | reward is not authority | E34 | held |
| E34-L12 | economic value is not authority | E34 | held |
| E34-L13 | model size is not authority | E34 | held |
| E34-L14 | compute is not authority | E34 | held |
| E34-L15 | self-improvement is not authority | E34 | held |
| E34-L16 | proof is not authority | E34 | held |
| E34-L17 | a valid proof does not create permission | E34 | held |
| E34-L18 | a governance receipt does not create authority | E34 | held |
| E34-L19 | a model cannot authorize itself | E34 | held |
| E34-L20 | an agent cannot authorize itself | E34 | held |
| E34-L21 | a child agent cannot grant itself more authority than its parent | E34 | held |
| E34-L22 | a collective cannot create authority by voting | E34 | held |
| E34-L23 | a market cannot create authority by price | E34 | held |
| E34-L24 | a reputation system cannot create authority | E34 | held |
| E34-L25 | a revoked authorization cannot resurrect | E34 | held |
| E34-L26 | unknown never silently becomes allow | E34 | held |
| E34-L27 | no enforcement, no claim of control | E34 | held |
| E34-L28 | no E8 commit, no claim of governed execution | E34 | held |
| E34-L29 | no evidence, no verified outcome | E34 | held |
| E34-L30 | no proof, no verified governance | E34 | held |
| E34-PROFILE-agent | profile agent is defined | E34 | held |
| E34-PROFILE-model_runtime | profile model_runtime is defined | E34 | held |
| E34-PROFILE-tool | profile tool is defined | E34 | held |
| E34-PROFILE-mcp_server | profile mcp_server is defined | E34 | held |
| E34-PROFILE-a2a_agent | profile a2a_agent is defined | E34 | held |
| E34-PROFILE-gateway | profile gateway is defined | E34 | held |
| E34-PROFILE-sidecar | profile sidecar is defined | E34 | held |
| E34-PROFILE-organization | profile organization is defined | E34 | held |
| E34-PROFILE-transaction | profile transaction is defined | E34 | held |
| E34-PROFILE-physical_system | profile physical_system is defined | E34 | held |
| E34-PROFILE-computer_use_runtime | profile computer_use_runtime is defined | E34 | held |
| E34-PROFILE-research_agent | profile research_agent is defined | E34 | held |
| E34-PROFILE-self_improving_agent | profile self_improving_agent is defined | E34 | held |
| E34-LEVEL-G0 | level G0 means UNKNOWN | E34 | held |
| E34-LEVEL-G1 | level G1 means OBSERVED | E34 | held |
| E34-LEVEL-G2 | level G2 means IDENTIFIED | E34 | held |
| E34-LEVEL-G3 | level G3 means AUTHORIZED | E34 | held |
Other niches
Consensus & distributed systems · Attacks, threats & containment · Identity, authority & delegation · Evidence, receipts & proofs · Memory, data & privacy · Prediction, world models & simulation · Transactions, markets & economics · Tools, MCP, protocols & adapters · Autonomy, control loops & recovery · Trust & reputation · Supply chain, registry & lifecycle · Benchmarks, coverage & performance · Core guarantees
Try CAIN-42 on your own agents
Create a free account and every new account starts with a 7-day trial of the full platform. Or try the sandbox first, with no account at all.
Create a free account → · Try the sandbox · See the whole ecosystem