Pull request: Add ERC: Agent Safety Framework (ASF) by garyyang-finchip · Pull Request #2113 · ethereum/ERCs · GitHub (ERC-8468, number assigned on 2026-10-27)
Repository: GitHub - garyyang-finchip/agent-safety-framework: AI Agent Safety Framework · GitHub (reference implementation, schemas, vectors, worked examples, two closed-loop consumers, design memo)
Current text: revision 10 (see the revision note below and the replies that introduce each revision). The revision-10 contracts are deployed and verified on Sepolia (third evidence run, 2026-11-15; see “Closing the loop”).
The gap
Agents now hold credentials, operate accounts and move funds for people and organizations, and the mechanisms that bound them are maturing in several places at once: permission grants and delegations in accounts, regulated mandates, action-bound policy verdicts, identity and reputation registries, decision mechanisms, task escrows. What none of them records is the relation between them: in this task, on these resources, in this execution environment, which participants have accepted which safety requirements, in which version, relying on which evidence, enforced by which mechanism, and whether that acceptance can still be relied upon now.
Without that relation a counterparty either trusts a label or asks for nothing; a requirement written into a document cannot be told apart from one enforced at an execution boundary; an accepted requirement is silently reinterpreted when its publisher releases a new version; and when an arrangement is paused or breached there is no shared way to say what stops, what continues, and who was entitled to intervene.
What the draft standardizes
Two objects, one check, one composition rule. Everything else is consumed from the standards that already provide it; the text requires only ERC-165.
Safety Specification. A versioned, content-addressed document: protected interests (five domains — assets and economic position, information and privacy, control and rules, credit and eligibility, human and public interests — each naming whose interest is protected: principal, operator, counterparty, bystander, public), applicability with permitted and prohibited uses, constraints under named profiles (evaluated over an action, a state, a trajectory, a group or a relation between trajectories), who may adopt and enforce and on what external authority, expected evidence and what it binds, enforcement channels with their location (self, account, runtime, custodian, contract, collective), coverage and known bypasses, failure handling for unknown / denied / dependency failure, pre-accepted remedies (reversibility, intervention window, service continuity, incident reporting), and the profiles a consumer must understand. Versions are immutable; publishing v2 changes nothing for adopters of v1; deprecation is a signal.
Safety Binding. The on-chain statement that an adopter (always msg.sender) adopts one pinned specification version for one subject within one scope, under a named external authority basis, with declared enforcement locations, a validity window, named interveners with explicit rights (suspend / resume / terminate) and a lifecycle: Active, Suspended, Terminated. A suspension records who made it, and a suspension recorded by an intervener is not the adopter’s to lift: only an intervener holding the resume right can lift it, while the adopter keeps the right to terminate (a terminated record keeps showing the stop it carried); and an intervener’s suspension is never blocked, neither by an existing suspension (it widens a record the adopter suspended itself and makes it the intervener’s, so the adopter cannot pre-empt a suspend-only intervener with a milder set of its own) nor by termination (it records the intervener’s stop through a terminated record, which itself never changes); a record’s response document is always its owner’s, and each intervener’s document is the document of its stop. An intervener’s suspension is also a standing stop against the adopter for the subject and the specification, keyed by the intervener, and the stop is the intervener’s: it outlives the termination of the binding it was recorded through, widens by union when the intervener suspends another binding of the adopter (a response document supplied with the widening states the widened requirements or names, in supersedes, the document it replaces as the stop’s reference), is lifted only by that intervener or by the one lifter it named (by no right of the adopter, and not by the resume holders of a binding), and no binding of that adopter for that subject and specification satisfies a clause that names the intervener while it stands — a sibling binding held in advance and a fresh adoption are refused alike. A stop, like a suspension, reaches the subject’s own actions under the binding, not the rights of counterparties over their own assets. A suspension carries a combinable set of response requirements (no new risk, shrink, essential only, de-risk only, full stop; full stop dominates) and the hash of a structured intervention response document that states who started the response and under which right, what it covers, what it prohibits, allows and keeps running, how in-flight actions are handled, how long it lasts and who may end it. The adopter is part of every identifier and index key, so the registry can never attribute an arrangement to an account that did not accept it; whether the adopter was entitled to adopt for the subject is verified by consumers against the permission source the binding references, never by the registry. Bindings can be derived from a parent (same pinned version, validity inside the parent’s, ancestors must hold, parameters never wider under the profile’s declared order).
The check. check(requirement, bindingIds) evaluates a requirement set (clauses with explicit version lists, adopter, subject, scope, required enforcement locations and required interveners — a resource can insist that someone it trusts is able to stop the subject, and a clause that names interveners also closes the resume right to the accounts it names, so the subject cannot lift a stop through an intervener of its own choosing — combined by all-of / any-of / k-of) and returns Satisfied, NotSatisfied or Unknown with a reason and the response set in force per clause. That set is the union over the binding, its suspended ancestors and the standing stops of the accounts the clause names (empty when the chain is terminated or expired), returned whatever reason is reported first, so a parent’s de-risk-only response reaches derived bindings intact, a full stop anywhere in the chain is visible below it, and a record the adopter suspended itself with a mild set cannot hide an intervener’s full stop from a consumer that applies the set. Pure registry logic — existence, state, validity, pin, adopter, subject, scope, enforcement, required interveners, standing stops by the accounts the clause names, ancestors — no policy is evaluated, Unknown never counts as satisfied, and clause verdicts are never combined into a score. The verdicts and reasons are the registry’s own; when a consumer also reads a permission, policy or validation mechanism, that mechanism’s native result is kept as it is, never normalized into the registry’s vocabulary. Which bindings a resource requires is decided by the resource’s own authority rules; a subject cannot substitute a looser one.
Closing the loop
Two reference consumers show that the relation bites at an execution boundary. ASFGuardedAccount: the operator may act only while the principal’s binding is satisfied; a guardian’s suspension with “no new risk” and “de-risk only” set blocks new exposure while repayment and cancellation still go through; once “full stop” is in the set, everything stops; there is no path that skips the check. ASFTaskTenderAdapter: an acceptance authority for token-bound task tenders that admits only fulfillers holding their own binding with the task publisher as a required intervener, rejects an unbound fulfiller and releases the reservation, rejects a fulfiller whose binding the publisher suspended (the fulfiller cannot lift it), and on Unknown neither accepts nor rejects — the kernel’s own judgment clock stays in force. Both are tested against negative scenarios (101 tests, 75 negative), including derived bindings on the wallet path, the closed resume right, the standing stops on both paths, the set in force under a self-suspension, the widening of a self-suspension by an intervener and a stop recorded through a terminated binding.
The evidence run has been executed on Sepolia three times, against the live TASK-KERNEL v3.0 TaskToken: with the revision-5 contracts on 2026-10-26 (36 transactions), with the revision-7 contracts on 2026-10-27 (43 transactions) and with the current revision-10 contracts on 2026-11-15 (blocks 12140959 to 12140990, 58 transactions, all successful, eight contracts verified on Sourcify with a runtime and a creation match). The third run carries the phases of the second (the guard loop, where the guardian’s suspension records its stop and its own resume lifts it; the tender loop; the standing-stop loop, where the task publisher stops the worker and the worker’s fresh binding is rejected with StoppedByIntervener until the publisher lifts its stop through liftStop; the lifecycle phase) and adds the set loop of revision 8: the guarded account’s clause names the guardian; the guardian suspends the presented binding with its de-risk-only response and widens its stop to a full stop through a sibling binding, with a document that names the one it supersedes; repayment through the account is refused although the presented record shows de-risk only, and refused again after the principal presents a fresh binding it suspended itself with de-risk only, because the guardian’s full stop is in the response set in force; once the guardian lifts its stop, repayment passes and borrowing does not until the principal resumes. Revision-10 addresses: ASFSpecificationRegistry 0x9cC8A10207F6B3E54241657cD040580939dDef5A, ASFBindingRegistry 0x17D5A20094d5C97C53d921b101Fe8A9eE7c031D7 (libraries ASFEvaluation 0x7Dad89304047AC97814405921d140616f1708032, ASFLifecycle 0x00706AD939f32e8e40833658485b410Def0662Eb), ASFGuardedAccount 0xb2E426d9b2A6f9A884ff9303288949267D76863D, ASFTaskTenderAdapter 0x2b7662128D9bc8Bd59E4a1206C19EE71be9A8cFf. The transaction records, the decoded events and the state read back after each run are in docs/sepolia-evidence-rev10.md, docs/sepolia-evidence-rev7.md and docs/sepolia-evidence.md of the repository; the earlier deployments stay as they were (their addresses are in the README).
What it deliberately does not do
It does not decide whether an action is permitted, does not evaluate policies, does not certify that an agent is aligned, and does not prove that an arrangement is enforced: declared locations are attributable commitments, evidence and the mechanisms carry the verification. Incident reports, claims and labels have no effect of their own in the framework; only the rules of the consumer that reads them give them one, and the text fixes the consumer rules that keep such records from being used as weapons. It records who accepted what, where, in which version, enforced where, and whether that is in force now — and makes that answer available in one call at the moment of effect.
Analysis basis
The design memo in the repository maps the problem space on three axes used as a multi-label index rather than as an exclusive classification: the seven links of an agent system (substrate, identity, authority, information flow, decision, accountability, collective), the five protected domains above, and six sources of failure. Eight requirement families group the profiles. The memo carries the object-and-authority attribution table, a scenario walkthrough against the test suite, and the overlap and adaptation review against the neighbouring proposals (permissions and delegations, policy verdicts, mandates, identity and reputation, validation, swarm charters, task tenders).
Questions for the thread
- Derived bindings: is the split between the on-chain structural rule and the off-chain profile-order rule acceptable, or should profiles with a declared order be checkable on chain?
- Response requirements: are the five combinable bits plus the structured response document the right shape, or should the response document’s
inFlightrule (what happens to actions already authorized or sent) also be reflected on chain? - Protected domains: the fifth domain, human and public interests with the roles principal / operator / counterparty / bystander / public, replaced a narrower “third parties and the public” domain. Does anyone see an interest that falls outside the five?
- Incident reports and claims stay outside the kernel and have effect only through consumer rules (and, if wanted, an optional profile). Is that the right boundary for a first version?
- Standing stops: an intervener’s stop applies to the adopter for the subject and the specification, keyed by the intervener, and is the intervener’s to lift (itself, or the one lifter it names with
setStopLifter). Is one lifter per intervener, for all its stops, the right granularity, or should a lifter be named per stop? And should a “created before” clause member still be offered for the replacement bindings that no stop concerns?
Feedback on the overlap review with the mandate, verdict, delegation and identity proposals is especially welcome.
Revision note
- Revision 1 (initial filing): two objects, one check, five exclusive suspension restrictions, four protected domains plus “third parties and the public”.
- Revision 2: suspension restrictions became a combinable bit set with a structured intervention response document (
responseRef); the fifth protected domain became human and public interests with protected roles; specifications gained permitted and prohibited uses and the remedies reversibility, intervention window, service continuity and incident reporting; the text now states that native results of external mechanisms are preserved rather than normalized, that identity, representation and legitimate use are established separately, and the consumer rules for incident reports, claims and labels; theIASFBindingRegistryinterface identifier changed accordingly (0xb99bdc2a); the design memo with the three-axis map, the attribution table, the scenario walkthrough and the overlap review was added to the repository. - Revision 3 (after the first review in this thread):
checkreturns the response set in force per clause, defined as the union over the binding and its suspended ancestors and empty when the chain is terminated or expired, and the reference account applies it forSuspendedandParentNotSatisfiedalike (a parent’s de-risk-only response no longer becomes a full stop on derived bindings); clauses gained required interveners (IntervenerMismatch); the registry recordssuspendedBy, and a suspension recorded by an intervener can be lifted only by an intervener holding the resume right, never by the adopter; the tender example requires the task publisher as an intervener; theIASFBindingRegistryinterface identifier is now0x1ff33befandrequirementIdvalues changed with the clause encoding; 82 tests, 56 negative. - Revision 4 (after the second review): a clause that names interveners also closes the resume right — the binding must not grant
RIGHT_RESUMEto any account the clause does not name, otherwise the adopter, who chooses the interveners, could lift any intervener’s suspension through an account of its own; the zero-account wildcard was removed for the same reason; additional suspend-only or terminate-only interveners remain allowed. No interface or encoding change; 84 tests, 58 negative. - Revision 5 (Sepolia): the evidence run was executed on Sepolia against the live TASK-KERNEL v3.0
TaskToken(36 transactions, all successful, seven contracts verified on Sourcify) and recorded in the repository. One reference-implementation change, made so that the registry could be deployed under Sepolia’s current gas schedule and its per-transaction gas cap: the binding registry’s check now runs in the externally linked libraryASFEvaluation, called with the registry’s storage. No change to the text’s normative content, the interfaces, the identifiers, the encodings or the test results. The title of the text is Agent Safety Framework (ASF), as in this thread’s title; revisions 2 to 4 of the pull request had carried “Agent Safety and Security Framework”, and the rationale paragraph on the name now explains the broad sense in which “safety” is used. - Revision 6 (after the third review): an intervener’s suspension is recorded as a standing stop against the adopter for the subject and the specification, keyed by the intervener; it outlives the termination of the binding it was recorded through, is lifted only through a
resumeby an account holding the resume right (also on the terminated record), and a clause that names the intervener is met by no binding of that adopter for that subject and specification while it stands (StoppedByIntervener, with the stop’s response set); terminated records keep their suspension fields;stopOfwas added and theIASFBindingRegistryinterface identifier is now0x2db6eb8a(the clause encoding andrequirementIdvalues are unchanged); the tender example names the publisher with the suspend and the resume right; the registry’s writes moved into a second linked library (ASFLifecycle); 90 tests, 64 negative. The Sepolia deployment remains revision 5’s. - Revision 7 (after the fourth review): the standing stop is the intervener’s. It is lifted only by the intervener that recorded it or by the one lifter the intervener named (
setStopLifter), throughliftStopor through that account’s ownresumeof a binding the intervener suspended; a resume holder on a binding reopens the record and nothing more, and aTerminatedrecord accepts noresume. A later suspension by the same intervener widens the stop by union (set, latest binding, latest response document supplied) instead of replacing it.StopRecorded,StopLifterSetandstopLifterOfwere added,StopLiftedis keyed by the stop, and theIASFBindingRegistryinterface identifier is now0xcff7f826(the clause encoding andrequirementIdvalues are unchanged). 93 tests, 67 negative. The evidence run was then executed on Sepolia a second time with the revision-7 contracts (2026-10-27, 43 transactions including the stop loop, all successful, eight contracts verified on Sourcify;docs/sepolia-evidence-rev7.md); the revision-5 deployment stays as it was. - Revision 8 (after two consumer reports on revision 7): the response set in force is defined for the clause, not for the record, and carries the standing stops of the accounts the clause names whatever reason the check reports first (
Suspended,ParentNotSatisfiedandStoppedByInterveneralike), so that a binding the adopter suspended itself with a mild set cannot hide an intervener’s full stop from a consumer that applies the set; the response document supplied with a suspension that widens a standing stop must state the widened requirements or name the documents it supersedes (supersedesmember of the response document), the stop’sresponseRefbeing the latest document while its set only widens; the text states that a stop reaches the subject’s own actions under the binding, not the rights of counterparties over their own assets. No interface or encoding change (0xcff7f826); the evidence script gained a seventh transaction phase for the set rule; 95 tests, 69 negative. The revision-7 deployment stays as it was; a third Sepolia run with the revision-8 contracts is prepared. - Revision 9 (after a third report): an intervener’s suspension is never blocked by an existing one. An intervener holding the suspend right may suspend a record that is already
Suspended: the record’s set becomes the union, its document is replaced when one is supplied, the record becomes the intervener’s when it carried the adopter’s own suspension, and the intervener’s stop is recorded as for any of its suspensions; the adopter’s own re-suspension stays invalid. This closes the pre-emption an adopter had against a suspend-only intervener (suspend one’s own binding first with a milder set, and the intervener’s escalation was an invalid transition with no stop ever recorded). No interface or encoding change (0xcff7f826); 98 tests, 72 negative. - Revision 10 (after the same reviewer’s probes of revision 9): a record’s response document is always the document of the account the record is attributed to, and only that account changes it (an intervener that widens a record carrying another intervener’s suspension writes only its own stop’s document; an intervener that takes over the adopter’s own suspension brings its own document, or none); and termination no longer blocks an intervener: its suspension of a
Terminatedrecord records its stop through it and leaves the record unchanged, so an adopter that terminates when it sees a suspension coming still meets the stop with a fresh binding. No interface or encoding change (0xcff7f826); 101 tests, 75 negative. Deployed and verified on Sepolia in the third evidence run (2026-11-15).