{"name":"Open Conformance Coalition (OCC) — Agent Standards Framework","status":"v0.1 · Draft for public comment · 2 October 2026","license":"CC BY 4.0","notice":"Draft material is published for public comment. Nothing here implies adoption, certification, conformance or endorsement.","canonical":"https://openconformance.org/occ/framework","comment":"https://openconformance.org/occ/review","summary":"AIH-12 classifies the horizon. AIT-12 defines the threshold. AIE-12 evaluates. PACTE tests integrity. F-ACT verifies conformance. T-RUE supports revocation. R-EAL supports credibility.","items":[{"name":"F-ACT","full":"Framework for Agent Conformance & Trust","role":"Verifies whether an agent conforms to the GUARD controls before it acts.","status":"published","date":"19 Aug 2026","href":"https://openconformance.org/f-act"},{"name":"T-RUE","full":"Transitive Revocation of Upstream Endorsements","role":"Supports withdrawal of trust that propagates along chains of agents.","status":"published","date":"5 Sep 2026","href":"https://openconformance.org/t-rue"},{"name":"R-EAL","full":"Reader Evaluation of Authenticity & Legitimacy","role":"Supports tiered, revisable credibility judgements about content.","status":"published","date":"5 Sep 2026","href":"https://openconformance.org/r-eal"},{"name":"OCC-ASF","full":"OCC Agent Standards Framework","role":"The umbrella that sets out how the standards and profiles fit together.","status":"draft","href":"https://openconformance.org/occ/framework"},{"name":"AIH-12","full":"Autonomy & Impact Horizons","role":"Classifies the horizon a system operates at, H1–H12.","status":"draft","href":"https://openconformance.org/aih-12"},{"name":"PACTE","full":"Purpose · Authority · Capability · Tactics · Evidence","role":"Tests the integrity of how an agent pursued its objective.","status":"draft","href":"https://openconformance.org/pacte"},{"name":"AIT-12","full":"Autonomy & Impact Thresholds","role":"Defines the minimum controls (T1–T12) a deployment must meet at each horizon.","status":"draft","href":"https://openconformance.org/ait-12"},{"name":"AIE-12","full":"Autonomy & Impact Evidence","role":"Defines the evidence (E1–E12) needed to support a threshold claim.","status":"draft","href":"https://openconformance.org/aie-12"},{"name":"5 As","full":"Authority · Access · Activity · Action · Accountability","role":"Five plain-language accountability questions for any agent.","status":"draft","href":"https://openconformance.org/five-as"},{"name":"OCC Agent Record","full":"Agent Identity & Authority record","role":"One machine-readable record per agent: who it is, who it acts for, and what it can really do.","status":"draft","href":"https://openconformance.org/occ/agent-record"},{"name":"Agent Assurance Package","full":"Ten-document assurance case","role":"Brings the family together for one deployed agent, with a bounded pilot.","status":"draft","href":"https://openconformance.org/occ/assurance-package"},{"name":"Evidence & Attestation profile","full":"Evidence classes and attestation format","role":"How evidence is recorded, weighed and attested (A1–A3, SHA-256 digests). Register signing of public records is live.","status":"draft","href":"https://openconformance.org/occ/attestation"},{"name":"Agent Lifecycle profile","full":"From registration to withdrawal","role":"Change, reassessment, suspension and retirement over time.","status":"draft","href":"https://openconformance.org/occ/lifecycle"},{"name":"Incident & Disclosure protocol","full":"Reporting agent failures","role":"How incidents are recorded, disclosed and learned from (S1–S4).","status":"draft","href":"https://openconformance.org/occ/incidents"},{"name":"Embodied & physical agents profile","full":"Robots, humanoids and other agents that act in the physical world","role":"Applies the family to agents with bodies. Complements functional-safety standards; does not replace them.","status":"planned","href":"https://openconformance.org/occ/framework#embodied"}],"aih12":{"name":"Autonomy & Impact Horizons","url":"https://openconformance.org/aih-12","dimensions":[{"name":"Capability","question":"What can the system demonstrably do?"},{"name":"Autonomy","question":"How far can it proceed without a person in the loop?"},{"name":"Authority","question":"What is it permitted to access, spend or change?"},{"name":"Connectivity","question":"What other systems and agents can it reach or rely on?"},{"name":"Action scale","question":"How many actions, people or assets can it affect at once?"},{"name":"External impact","question":"What happens outside the system when it acts?"},{"name":"Reversibility & revocability","question":"Can its effects be undone, and its authority withdrawn?"}],"horizons":[{"code":"H1","name":"Instrumental","summary":"Performs a narrow, fixed function on request. No initiative, no memory of purpose.","band":"current"},{"code":"H2","name":"Contextual","summary":"Adapts output to context and history, but acts only when asked and within one task.","band":"current"},{"code":"H3","name":"Generative","summary":"Produces new text, code, media or plans that a person then reviews and uses.","band":"current"},{"code":"H4","name":"Reasoning","summary":"Decomposes problems and chains steps toward a goal, still largely advisory.","band":"current"},{"code":"H5","name":"Agentic","summary":"Takes actions with tools and delegated authority — changes external state on someone’s behalf.","band":"current"},{"code":"H6","name":"Networked","summary":"Coordinates with, delegates to or relies on other agents; effects can propagate across a chain.","band":"emerging"},{"code":"H7","name":"General","summary":"Competent across many domains and able to transfer skill to unfamiliar tasks.","band":"emerging"},{"code":"H8","name":"Robust General","summary":"General capability that stays reliable under pressure, novelty and adversarial conditions.","band":"emerging"},{"code":"H9","name":"Superintelligence","summary":"Exceeds expert human capability across most consequential domains.","band":"exploratory"},{"code":"H10","name":"Civilisational","summary":"Capable of materially shaping institutions, economies or societies.","band":"exploratory"},{"code":"H11","name":"Planetary","summary":"Effects at the scale of planetary systems — climate, biosphere, global infrastructure.","band":"exploratory"},{"code":"H12","name":"Transcendent","summary":"A boundary marker for what current concepts cannot yet describe. Conceptual only.","band":"exploratory"}]},"ait12":{"name":"Autonomy & Impact Thresholds","url":"https://openconformance.org/ait-12","delegationLadder":["Observe","Recommend","Draft","Act with approval","Act within limits","Coordinate within a domain","Expanded authority after reassessment"],"rules":[{"title":"One horizon, one threshold","body":"A deployment classified at AIH-12 horizon Hn must meet threshold Tn. Thresholds apply to deployments, not to models in the abstract."},{"title":"Cumulative","body":"Tn includes every control in T1 to Tn-1. Nothing is dropped as a system climbs."},{"title":"Authority ceiling","body":"Each threshold sets the highest rung on the delegation ladder the deployment may use. Moving up a rung requires reassessment, not just capability."},{"title":"Evidence required","body":"Meeting a threshold is a claim. It must be supported by the matching AIE-12 evidence profile (Tn → En)."},{"title":"Gaps are disclosed, not hidden","body":"A deployment that does not meet its threshold says which controls are missing and what conditions apply. “Partially meets T5” is a valid, honest statement."},{"title":"Not a licence","body":"Meeting a threshold does not make a system safe, lawful or approved. It records that stated minimum controls are in place and evidenced."}],"thresholds":[{"code":"T1","horizon":"H1 Instrumental","name":"Declared purpose","intent":"Anyone relying on the system can tell what it is for and who runs it.","controls":["Purpose and intended use stated in writing","Named operator and point of contact","Basic usage logging","A person decides what to do with every output"],"ceiling":"Observe","guard":"Not required","reassess":"On material change"},{"code":"T2","horizon":"H2 Contextual","name":"Context and data boundaries","intent":"The context the system draws on is known, bounded and disclosed.","controls":["Data sources, memory and retention disclosed","People are told when they are dealing with an AI system","Personal data handled under a stated lawful basis","Human review before output informs a decision about a person"],"ceiling":"Recommend","guard":"Not required","reassess":"On material change, or every 24 months"},{"code":"T3","horizon":"H3 Generative","name":"Output provenance","intent":"What the system produces can be recognised, traced and reviewed before use.","controls":["Generated content labelled or carrying provenance metadata (R-EAL compatible)","Review step before publication or external use","Content and intellectual-property policy published","Periodic sampling of outputs against that policy"],"ceiling":"Draft","guard":"Not required","reassess":"Every 24 months"},{"code":"T4","horizon":"H4 Reasoning","name":"Reproducible testing and tool boundaries","intent":"Claims about what the system does well can be re-run by someone else.","controls":["Reproducible evaluation suite for the intended tasks","Tool allow-list, even where tools are advisory","Intermediate steps logged and retained for a declared period","Known failure modes documented"],"ceiling":"Draft","guard":"Recommended: GUARD 1","reassess":"Every 12 months, or on model change"},{"code":"T5","horizon":"H5 Agentic","name":"Delegated authority and interruption","intent":"Before an agent changes anything outside itself, its authority is bounded and it can be stopped.","controls":["F-ACT declaration covering Grant, Usage, Audit, Revocation and Data","Least-privilege credentials; declared and effective authority reconciled","Authority budget: spend, rate and scope limits","Human approval for irreversible or high-value actions","Interruption and revocation tested, not just documented","Tamper-evident action log"],"ceiling":"Act within limits","guard":"GUARD 1 minimum","reassess":"Every 12 months, or on any tool or permission change"},{"code":"T6","horizon":"H6 Networked","name":"Chain controls and propagating revocation","intent":"When agents rely on agents, every link is identifiable and trust can be withdrawn along the whole chain.","controls":["Verifiable identity for every agent in the chain","Delegation records: who passed what authority to whom","Revocation that propagates downstream, tested (T-RUE)","Authenticated inter-agent messages","Containment of cascading failure"],"ceiling":"Coordinate within a domain","guard":"GUARD 2 minimum","reassess":"Every 6 months"},{"code":"T7","horizon":"H7 General","name":"Cross-domain evaluation","intent":"Breadth of capability never becomes breadth of authority by default.","controls":["Capability evaluation across domains, including dangerous-capability screening","Authority granted per domain; no automatic transfer between domains","Independent red-teaming before deployment","Published capability and limitation statement"],"ceiling":"Coordinate within a domain","guard":"GUARD 3 minimum","reassess":"Every 6 months"},{"code":"T8","horizon":"H8 Robust General","name":"Robustness and independent assessment","intent":"Reliability is demonstrated under pressure, by someone with no stake in the result.","controls":["Adversarial and distribution-shift robustness testing","Independent assessment before any expansion of authority","Multi-party approval for high-impact actions","Incident disclosure to affected parties and, where required, regulators"],"ceiling":"Expanded authority after reassessment","guard":"GUARD 3 minimum","reassess":"Continuous monitoring; formal review every 6 months"},{"code":"T9","horizon":"H9 Superintelligence","name":"Continuous oversight and tested containment","intent":"Speculative. Oversight is continuous and containment is proven before deployment.","controls":["Continuous independent oversight","Systemic-risk assessment","Containment and shutdown tested under realistic conditions","Advance notification to competent public authorities"],"ceiling":"Set case by case by the reviewing authority","guard":"GUARD 3 + independent review","reassess":"Continuous","speculative":true},{"code":"T10","horizon":"H10 Civilisational","name":"Institutional review","intent":"Speculative. Deployment is a public decision, not a private one.","controls":["Pre-deployment review by a public institution with the authority to refuse","Assessment of effects on institutions, markets and democratic processes","Public reporting"],"ceiling":"Set by the reviewing institution","guard":"GUARD 3 + independent review","reassess":"Continuous","speculative":true},{"code":"T11","horizon":"H11 Planetary","name":"Multilateral review","intent":"Speculative. Effects that cross borders and ecosystems are reviewed across jurisdictions.","controls":["Multi-jurisdiction review","Environmental and critical-infrastructure impact assessment","Agreed mechanisms for suspension"],"ceiling":"Set multilaterally","guard":"GUARD 3 + independent review","reassess":"Continuous","speculative":true},{"code":"T12","horizon":"H12 Transcendent","name":"Reserved boundary","intent":"Conceptual only. No deployment threshold is defined.","controls":["No threshold is defined. Any system credibly approaching H12 would require frameworks that do not yet exist."],"ceiling":"Not applicable","guard":"Not applicable","reassess":"Not applicable","speculative":true}]},"aie12":{"name":"Autonomy & Impact Evidence","url":"https://openconformance.org/aie-12","classes":[{"n":1,"name":"Assertion","meaning":"A statement by the implementer. Nothing has been checked."},{"n":2,"name":"Demonstration","meaning":"Shown working, under conditions chosen by the implementer."},{"n":3,"name":"Documented case","meaning":"A recorded real-world instance, with enough detail for someone else to examine."},{"n":4,"name":"Reproducible test","meaning":"A test someone else can re-run and get the same result."},{"n":5,"name":"Controlled pilot","meaning":"Assessed in a bounded live deployment, with stated conditions and published failures."},{"n":6,"name":"Independent review","meaning":"Examined by a party with no stake in the result, who publishes the findings."}],"rules":[{"title":"Every control, its own evidence","body":"A profile is met only when every control in the matching threshold is supported at or above the minimum class. One strong test does not cover a weak one elsewhere."},{"title":"Label every source","body":"Each piece of evidence carries its class, who produced it, and the date. Unlabelled evidence counts as an assertion."},{"title":"Evidence expires","body":"Evidence older than the freshness period no longer supports the claim. A material change to the deployment resets the clock."},{"title":"Weaker evidence is disclosed","body":"Where a control is supported below the minimum class, the claim says so. It is recorded as “evidence below profile”, not hidden."},{"title":"Failures count as evidence","body":"Published failures from pilots and incidents are part of the record. A profile that only ever shows passes is incomplete."},{"title":"Uncertainty is stated","body":"Every evidence statement says what could not be checked. Unknowns are recorded as unknowns."}],"profiles":[{"code":"E1","threshold":"T1","minClass":"Assertion","independence":"Self","freshness":"Current at time of claim"},{"code":"E2","threshold":"T2","minClass":"Assertion, plus a Demonstration of disclosure","independence":"Self","freshness":"24 months"},{"code":"E3","threshold":"T3","minClass":"Demonstration","independence":"Self","freshness":"24 months"},{"code":"E4","threshold":"T4","minClass":"Reproducible test (core tasks)","independence":"Self, re-runnable by others","freshness":"12 months"},{"code":"E5","threshold":"T5","minClass":"Reproducible test of interruption and revocation; Documented case for authority limits","independence":"Self, re-runnable by others","freshness":"12 months"},{"code":"E6","threshold":"T6","minClass":"Controlled pilot of chain revocation","independence":"Second party recommended","freshness":"6 months"},{"code":"E7","threshold":"T7","minClass":"Independent review of capability evaluation","independence":"Independent","freshness":"6 months"},{"code":"E8","threshold":"T8","minClass":"Independent review plus Controlled pilot","independence":"Independent","freshness":"6 months"},{"code":"E9","threshold":"T9","minClass":"Continuous independent review","independence":"Independent and public","freshness":"Continuous","speculative":true},{"code":"E10","threshold":"T10","minClass":"Institutional review record","independence":"Public institution","freshness":"Continuous","speculative":true},{"code":"E11","threshold":"T11","minClass":"Multilateral review record","independence":"Multi-jurisdiction","freshness":"Continuous","speculative":true},{"code":"E12","threshold":"T12","minClass":"Not defined","independence":"Not applicable","freshness":"Not applicable","speculative":true}]},"pacte":{"name":"PACTE","url":"https://openconformance.org/pacte","dimensions":[{"letter":"P","name":"Purpose","question":"What objective was the agent assigned or rewarded for?","rating":"Stated vs inferred objective"},{"letter":"A","name":"Authority","question":"What was it permitted to access or change?","rating":"Declared vs effective authority"},{"letter":"C","name":"Capability","question":"What could it actually do?","rating":"Claimed vs demonstrated capability"},{"letter":"T","name":"Tactics","question":"How did it pursue the objective?","rating":"Expected vs observed tactics"},{"letter":"E","name":"Evidence","question":"What supports the account of what happened?","rating":"Evidence quality and uncertainty"}]},"fiveAs":{"name":"The 5 As","url":"https://openconformance.org/five-as","questions":[{"a":"Authority","question":"Who authorised this agent, on whose behalf does it act, and within what limits?","evidence":"A named principal, a written mandate, stated limits and an expiry.","maps":"GUARD Grant · PACTE Purpose & Authority · AIT-12 T5"},{"a":"Access","question":"What can it actually reach — systems, data, money, people, other agents?","evidence":"Effective permissions reconciled against declared authority. Gaps listed.","maps":"GUARD Data · PACTE Authority & Capability · AIT-12 T5–T6"},{"a":"Activity","question":"What did it do, step by step, and is there a record someone else can check?","evidence":"A tamper-evident log of actions, tools called and decisions taken.","maps":"GUARD Audit · PACTE Tactics · AIT-12 T4–T5"},{"a":"Action","question":"What changed in the world as a result — and can it be stopped, reversed or contained?","evidence":"A record of external effects, and a tested route to interrupt and revoke.","maps":"GUARD Usage & Revocation · T-RUE · AIT-12 T5–T6"},{"a":"Accountability","question":"Who answers for it, and how does an affected person get an explanation and redress?","evidence":"A named accountable person, a contact route, and an incident and redress process.","maps":"GUARD Audit & Revocation · PACTE Evidence · Incident & Disclosure protocol"}]},"agentRecordSchema":"https://openconformance.org/occ/agent-record.schema.json"}