HDP

Autonomous AI agents increasingly execute consequential actions: sending email, modifying files, running code, calling APIs, and transacting on behalf of people. When a human authorizes an orchestrator agent, which delegates to sub-agents, which delegate again to tool-execution agents, the originating human authorization becomes disconnected from the terminal action. There is no standard record of the authorization chain.

That gap creates three concrete problems:

HDP addresses this by defining a token that:

HDP is not an authorization protocol, and an HDP token is not a capability, an access token, or a credential that entitles its holder to anything. Presenting a valid HDP token to a service does not authorize the presenter to perform the requested action. That decision belongs to the service’s own access control mechanism, whether that is OAuth 2.0 (§12.2), a capability system such as UCAN or ZCAP-LD (§12.4, §12.5), or something else. HDP is designed to travel alongside such mechanisms, not to replace them.

Several parts of this document can be misread as authorization if this distinction is not kept in view. The scope object (§3.3) records what the human declared, in fields named authorized_tools and authorized_resources; it is a signed record of the authorization event, not a grant. The verification pipeline (§5) establishes that a token is authentic and intact, not that its presenter may act. The HTTP transport (§8) shows a token accompanying a request because the task travels in the request, not because the token authorizes it.

The primary reader of an HDP token is therefore not the service receiving a request but whoever examines the record afterwards: post-incident reconstruction of which agent did what, under whose authorization, and in what order; compliance evidence that a human authorized a class of action; and human oversight, where an approver inspects the chain a task has accumulated before permitting it to continue. A token is carried at invocation and read at audit.

The need for agentic delegation provenance is not hypothetical. Production deployments of AI orchestration systems, including LangChain, AutoGPT, CrewAI, and similar frameworks, pass natural language task descriptions between agents with no cryptographic binding to the original human authorization. The operational risk compounds as models become more capable and agents receive access to higher-consequence tools.

A provenance token that travels alongside the task, tamper-evident, offline-verifiable, and scoped to what the human actually approved, provides the foundation for auditable, accountable agentic systems.

HDP is designed with the following goals in order of priority:

The Intent Provenance Protocol (IPP) addresses the same problem space. HDP and IPP share the use of Ed25519 signatures and append-only provenance chains but make different architectural trade-offs, which are detailed in §12.1. The two protocols are not interoperable. HDP is offered as a distinct design point, not a revision of IPP.

The full HDP protocol specification is available at [HDP-SPEC]. A TypeScript reference implementation is available at [HDP-IMPL].

The core of HDP is an append-only, cryptographically chained record: each hop extends a signed entry that covers all prior state, gaps in the hop sequence are tamper-evident, and any party can verify the entire chain offline using only a public key. This chain-of-custody mechanism is independent of what the chain carries.

This document profiles that mechanism for one application: human-authorized agentic delegation. In this profile the carried payload is the scope object (§3.3) and each hop record describes an agent delegation action. The same mechanism could carry other payloads, for example data provenance, consent delegation, or physical-world command chains, each as a distinct profile. Such profiles are out of scope for this document; HDP v0.1 defines only the agentic-delegation profile. Where practical, the signing (§4, §4.1, §4.2) and verification (§5) procedures are described in a payload-agnostic way so future profiles can reuse them unchanged.

The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, NOT RECOMMENDED, MAY, and OPTIONAL in this document are to be interpreted as described in RFC 2119 and RFC 8174 when, and only when, they appear in all capitals.

An HDP token is a JSON object with six top-level fields. The token MUST conform to the structure below. All integer timestamps are Unix milliseconds, that is, milliseconds since 1970-01-01T00:00:00Z.

Before signing or verifying a token, implementations MUST validate its JSON representation and all REQUIRED fields, types, and constraints defined in this section. The input MUST satisfy the I-JSON requirements of RFC 8785, including rejection of duplicate object member names, invalid Unicode strings, and non-finite numbers. Duplicate names MUST be detected before parsing discards them. Values MUST NOT be coerced from strings or booleans to satisfy a numeric field’s type.

The integer fields header.issued_at, header.expires_at, and each hop’s timestamp and parent_hop MUST be in the inclusive range 0 through 9007199254740991, that is 2⁵³ − 1. Each hop’s seq, and scope.max_hops where present, MUST be in the inclusive range 1 through 9007199254740991. These are JSON numbers, not strings. Implementations MUST check the numeric value before any lossy conversion, and MUST reject fractional or out-of-range values rather than round them. The bounds ensure exact integer representation in the IEEE 754 double-precision model used by RFC 8785. They are representation bounds, not an operational delegation budget. Other numeric values, such as numbers insideprincipal.metadata, remain subject to RFC 8785.

The header object carries token lifecycle and session binding fields.

The principal object identifies the authorizing human. It MUST contain id and id_type. All other fields are OPTIONAL.

Defined id_type values:

The scope object records what the human authorized. It is signed as part of the root signature and MUST NOT be modified after issuance.

The scope object is a record, not a grant. Its fields describe the authorization the human gave at issuance, so that the record can later be compared with what agents declared they did. Nothing in this object confers authority on an agent that holds the token. See §1.1.

The chain array is append-only. Each element records a single delegation event, called a hop. The array is empty at issuance and grows as the token passes between agents. Agents MUST NOT remove or modify existing entries.

The signature object carries the root signature computed by the issuer.

HDP uses Ed25519 (RFC 8032) over payloads serialized with the JSON Canonicalization Scheme (RFC 8785), with signatures encoded as base64url without padding (RFC 4648). Canonicalization is what makes the signed bytes deterministic across implementations and platforms.

The root signature is computed by the issuer at token creation time. It covers the fields that constitute the human authorization event: the header, principal, and scope.

Each hop MUST carry a hop_signature. It binds the new hop record to the entire accumulated delegation history and to the root signature, which makes retroactive chain modification detectable.

In HDP v0.1 all signatures, the root signature and every hop signature, are produced by the issuer using a single key. An extending agent that is not the issuer submits its hop to the issuer, which signs it and returns the extended token. Two consequences follow, and implementers should weigh both.

First, a v0.1 hop signature attests that the issuer recorded a delegation claim naming the agent in agent_id. It does not attest that the named agent consented to, or knew of, the hop, because the agent signed nothing. The chain is a record of what the issuer recorded, not of what each agent agreed to. Deployments in which that distinction matters need per-agent signing.

Second, the single-key design is practical only where the issuer is reachable whenever any agent wishes to extend the chain. This adds a round trip to every delegation, and it makes delegation across trust domains awkward, since an issuer in one domain must sign on behalf of agents in another. Only verification is offline; extension is not.

The single-key design does not, however, gain anything for offline verification that per-agent signing would lose. If each hop carried the public key of the agent appending it, signed into the chain by that agent’s delegator, a verifier would authenticate every key after the first from the chain itself and would still resolve exactly one key out of band: the issuer’s. Per-agent hop signing on that pattern is the planned extension for a future version. It is not part of v0.1, and v0.1 tokens carry no per-agent keys.

These rules govern chain construction and MUST be enforced by both extenders and verifiers.

For live acceptance, a verifier MUST first validate the input as specified in §3, then execute the following seven steps in order. A failure at any step MUST cause immediate rejection for live use with an appropriate error, and the live acceptance pipeline MUST NOT proceed to subsequent steps. Historical audit is a separate procedure, defined in §5.1. Rejecting live use does not prohibit that procedure from examining the same record.

An optional eighth step MAY be performed where the application has registered a Proof-of-Humanity verifier: if principal.poh_credential is present and a verifier callback is configured, the credential MUST be validated by that callback, and the token MUST be rejected if validation fails. The callback runs last because it is application-defined and may be remote, costly, or side-effecting. Ordering it after Steps 1 through 7 keeps those steps offline and ensures that no external verifier is invoked for a token that fails its cryptographic checks. See §9.3.

A token that passes all seven steps is authentic and intact: its header, principal, and scope are as the issuer signed them, and every recorded hop is as the issuer recorded it. Passing verification establishes nothing about whether the presenter may perform any action. That determination is made by the application’s own authorization mechanism, to which the verified token is an input. See §1.1.

An auditor MUST be able to examine an expired or revoked token without treating it as acceptable for a new request. Audit tooling MUST report the following three results separately. They are verification results, not new fields in an HDP token.

Historical evidence SHOULD include an authenticated receipt or integrity-protected verifier log binding the complete token digest (§8.2), the observed request or event, the session identifier, the verifier identity, the time, the decision, and the relevant revocation and policy state. Any required Proof-of-Humanity result belongs in that evidence; a credential check performed today is not evidence of its status then. A declared hop timestamp alone is not trusted time evidence, and an empty current revocation set does not establish past status. These records are application-layer artifacts whose encoding is outside HDP v0.1, and they can be retained and checked offline.

Long-running or streaming sessions may exhaust the max_hops limit, require scope expansion, or reach a high-risk action that warrants fresh human confirmation. In these cases the issuer, acting on behalf of the human principal, issues a new token that supersedes the original.

Re-authorization is indicated by setting header.parent_token_id to the token_id of the token being superseded. This field MUST be set before computing the root signature, so the parentage link is cryptographically covered by the new token’s root signature.

A re-authorized token:

Verifiers that require re-authorization chain traversal SHOULD retain all tokens in a session and verify the full parent_token_id linkage.

HDP v0.1 supports one principal per token. Joint authorization by multiple humans is achieved by sequential chaining: Human A issues token T1, then Human B issues token T2 with parent_token_id equal to T1’s token_id. Each token is independently signed with its issuer’s key. To verify a multi-principal chain, a verifier MUST:

This provides auditable joint authorization without a threshold signature scheme. Each principal’s authorization is a distinct signed artifact. Simultaneous multi-signature primitives using threshold schemes are planned for a future version.

An alternative to chaining is composition: each principal issues an independent token, and the verifier’s policy requires that both be presented. Composition is more general and composes further downstream, and a verifier MAY adopt it. Chaining is specified here because T2 signs a reference to T1, making the linkage part of the record. The link alone does not state that the authorizations were conjunctive or approved the same action; that meaning requires the application context described below. Composition can also provide auditable evidence when an authenticated receipt binds both token digests to the request and to the policy requiring them. Without such retained context, neither a bare parent link nor two independent tokens establishes joint approval.

The HTTP header field names defined below do not use the X- prefix, in accordance with RFC 6648.

Tokens MAY be transmitted in HTTP requests and responses using the HDP-Token header. The value is the base64url encoding, no padding, of the UTF-8 JSON serialization of the complete token object.

A token accompanies a request because the task it records travels in that request. Its presence does not authorize the request (§1.1). The receiving service decides whether to act by its own means and MAY use the verified token as an input to that decision.

HTTP header values are routinely written to access logs, proxy logs, and error reports, and the HDP-Token value contains the principal object and the full chain. Deployments SHOULD configure logging to treat HDP-Token as sensitive, as they would an Authorization header.

Where token size is a concern, for example with large chains, the token MAY be stored server-side and referenced using the HDP-Token-Ref header. The reference is either the token’s token_id, or a content-addressed reference: the string sha256: followed by the base64url encoding, no padding, of the SHA-256 digest (RFC 6234) of the token’s canonical JSON serialization per RFC 8785.

A recipient resolving a content-addressed reference MUST validate the resolved token’s input representation, serialize the complete token, including signature and all hop signatures, with RFC 8785, encode the result as UTF-8, compute SHA-256, and compare the digest with the reference. The digest MUST be exactly 32 bytes encoded as canonical unpadded base64url; a malformed encoding or a digest mismatch MUST cause rejection. The comparison commits to canonical JSON, not to whitespace or member ordering in the stored serialization. For a UUID reference, the resolved header.token_id MUST identify the same UUID, compared by its 128-bit value rather than hexadecimal letter case; a mismatch MUST be rejected.

Successful reference resolution MUST be followed by live acceptance or historical audit verification as appropriate. A valid signature alone does not establish that the requested reference was resolved correctly. Implementations using token-by-reference MUST secure the token store and use TLS for all reference resolution.

Issuers that wish to publish their Ed25519 public keys for automated discovery SHOULD serve a JSON document at /.well-known/hdp-keys.json with the following structure.

This endpoint is a discovery convenience only. It is not part of verification, which takes the issuer’s public key as an input and remains fully offline (§10.11). A verifier that has obtained the key by other means has no reason to consult it.

The format is intentionally minimal and implementations MAY extend it with additional metadata. The alg field MUST be Ed25519 for v0.1 keys. Consumers MUST reject entries with unrecognized alg values, and MUST validate that the decoded public key is exactly 32 bytes.

The principal object may contain PII, including an email address or display name. Issuers SHOULD apply the principle of minimum disclosure when constructing tokens that will traverse multiple agents.

The same principle applies to the agent_id field in hop records (§3.4). An agent_id need not be meaningful to anyone but the delegator that assigned it. This is sufficient because accountability in a delegation chain is recursive: a delegator is responsible for how its direct delegate uses the delegation, even when the use occurred further down the chain, and that delegate is in turn responsible for its own direct delegate. A verifier or auditor therefore never needs to resolve an agent_id globally. It needs the delegator at each step to be able to identify the party it delegated to, and the agent_id and its action_summary are bound into the signed chain for exactly that purpose.

Which identifier to use is context dependent, and HDP does not prescribe one. Non-exhaustively, a widely known identifier, such as an enterprise employee or service number, suits deployments where correlation is not a concern; an identifier meaningful only to the delegator suits deployments where an observer must be prevented from correlating requests across chains; a DID (W3C DID) suits deployments where a trusted authority exists to assert claims about it. An issuer or extending agent MAY use a fresh identifier for every delegation.

HDP v0.1 offers no field-level confidentiality: a token is either presented whole for verification or stripped and marked audit-only, as above. Encrypting principal fields is not specified, because the keys in circulation are Ed25519 signing keys rather than encryption keys and because the set of verifiers is deliberately open, so there is no defined party to encrypt to. Selective disclosure of principal fields and of individual hops, which would allow a token to be verified with parts withheld, is planned for a future version.

HDP tokens may constitute personal data under applicable privacy regulations, such as GDPR Article 4(1), when the principal.id or principal.display_name fields contain directly or indirectly identifying information.

Implementations SHOULD:

Encrypting stored tokens does not by itself discharge an erasure obligation, since the ciphertext remains personal data for as long as the key exists. Destroying the key, crypto-shredding, is a recognised technique for rendering retained tokens unreadable and MAY be used together with the mapping-destruction approach above.

The optional principal.poh_credential field MAY carry a credential attesting that the principal is a human, for example a Worldcoin World ID proof, a CAPTCHA session token, or a biometric attestation identifier. The HDP protocol does not define the semantics of this field; verification is entirely application-defined.

When a PoH verifier is configured, the verification pipeline MUST validate the credential as the final step, after session binding, and MUST reject the token if validation fails. The verifier callback SHOULD be idempotent and SHOULD NOT have side effects. §5 explains why the callback is ordered last.

HDP is designed to provide provenance and tamper evidence, not runtime enforcement. An agent that exceeds its declared scope is still a bad actor; HDP creates an evidence trail, not a capability boundary. Applications requiring runtime enforcement MUST implement it at the application layer using the HDP token as audit input. HDP is not an authorization protocol (§1.1); the properties discussed below are properties of the record: who could have produced it, whether it has been altered, and what it does and does not establish.

A forged token, one whose header, principal, or scope fields do not match the original issuance, will fail §5, step 3, the root signature check. The security of this step relies on the unforgeability of Ed25519 signatures and the collision resistance of SHA-512, used internally by Ed25519. An attacker that does not possess the issuer’s private key cannot produce a valid root signature for a modified token.

Modification, reordering, or removal of any non-trailing hop is detectable: it either breaks the hop sequence check (§5, step 4) or invalidates the hop signatures of all subsequent hops (step 5), because each hop signature covers all previous hops and the root signature. Insertion of a fabricated hop similarly fails unless the attacker possesses the issuer’s private key. Removal of one or more trailing hops is a distinct case that these checks do not detect; see §10.4.

Each hop signature covers only the hops that precede it and the root signature. Consequently, deleting one or more hops from the end of the chain, or presenting an earlier and shorter copy of a token, yields a token that still passes every step of the verification pipeline. HDP therefore provides tamper evidence for the hops that are present, but does not by itself prove that the chain is complete.

Relatedly, a non-cooperating or compromised agent can decline to append a hop for an action it takes. HDP records declared delegation actions and cannot compel an agent to record one. HDP is an evidence trail, not an enforcement mechanism (§10.1).

A verifier that can authenticate the presenter, for example because the transport identifies the calling agent, SHOULD require that the final hop’s agent_id correspond to that presenter. This is cheap and closes one truncation case: a third party holding a shorter, earlier copy of the token cannot present it, because the final hop of that copy names someone else. Capability systems impose the same requirement; UCAN Invocation requires the delegation chain to end at the invoker, and a ZCAP-LD invocation proof is rooted in the invoker’s key. See §12.4 and §12.5.

The presenter check does not make the chain complete. In a capability chain, truncation gains an attacker nothing beyond what the presenter check catches, because authority only narrows toward the tail: a truncated prefix is usable only by the delegatee of its last remaining hop, who holds that authority legitimately. HDP hops record actions, not grants, and there is no attenuation; a truncated chain carries the full original scope. The case HDP must consider is an intermediate agent that deletes the hops appended after its own, in order to hide what its sub-agents did. After truncation that agent genuinely is the presenter, and the presenter check passes.

Concurrent extensions from the same prefix can produce two valid branches with the same token_id, hop count, and final agent_id, but different recorded actions. A hop count or presenter check cannot distinguish them. The signature pipeline verifies the supplied branch; it does not discover other branches or select a uniquely final one.

Deployments that require one linear record per token MUST serialize extensions at the issuer. The issuer MUST atomically check that the submitted prefix matches its accepted chain head and advance that head when committing an extension. A stale prefix MUST be rejected for reconciliation against the current head. An already signed hop cannot simply be transplanted onto another branch; extending the reconciled prefix requires a new signature. Retries SHOULD return an already committed result when the application identifies the same extension request. Deployments that intentionally allow branching MUST retain and identify the branches separately and define how their audit process accounts for them. HDP v0.1 defines no branch merge operation. Issuer serialization is state used for construction, not a network dependency of verification.

Applications requiring evidence of a particular observed or final record SHOULD retain an authenticated receipt or settlement record binding the token_id, session_id, complete token digest as defined in §8.2, observation time, and observing party. It MUST distinguish an observed snapshot from a claimed final record. A verifier relying on that receipt MUST validate its authenticity and compare the token digest. A hop count MAY be included for diagnostics but MUST NOT be treated as a substitute for the digest. The receipt establishes which branch was observed or finalized under the application’s policy; it does not prove that no unrecorded action or undisclosed branch exists. Off-record delegation remains a separate limitation (§10.7). These receipts are application-layer artifacts, not new token fields.

HDP provides two orthogonal replay defenses:

Together, these defenses ensure that a stolen token is useful to an attacker only within the original session and only until it expires or is revoked (§10.6).

HDP specifies no default lifetime, deliberately. Expiry alone forces a choice between tokens that lapse just before they are needed and tokens that outlive a detected compromise, and the tendency in deployed systems is for lifetimes to lengthen over time as the first kind of failure accumulates operational friction. Issuers SHOULD choose the shortest lifetime the task permits, and SHOULD rely on revocation rather than on long lifetimes to avoid disruption, since revocation is what bounds the exposure of a compromised token regardless of the lifetime it was issued with.

Session binding says nothing about replay of the same token within its session. Whether a presenter may present one token for two requests is an application-layer question on which HDP takes no position. Applications for which it matters SHOULD record the requests each token has accompanied, for example by retaining a hash of each request until the token expires, and reject repeats.

Because session_id anchors the session-binding defense, it SHOULD be unguessable. Issuers SHOULD generate session_id values with at least 128 bits of entropy from a cryptographically secure random source. A predictable session_id weakens replay protection.

A verifier MUST support being instructed to stop honouring a token. The instruction identifies the token by header.token_id; the verifier records that identifier in local revocation state and thereafter rejects the token at §5, step 2. Revocation MUST NOT prevent separate historical integrity verification (§5.1). No central registry, no publication mechanism, and no network access at verification time are involved. The revocation state is held by the verifier, as the session_id already is, and verification remains fully offline (§10.11).

HDP does not specify who may revoke, how the instruction reaches the verifier, or the retention period for revocation evidence; these are verifier policy. Retaining an entry until the corresponding token’s expires_at is sufficient for live rejection, since the token is rejected on expiry thereafter. Historical audit may require longer retention of revocation events and effective times, as described in §5.1. The range of reasonable policies is illustrated by existing capability systems: UCAN allows a delegator to revoke and makes the right to revoke itself delegable, and ZCAP-LD allows any delegator in a chain to revoke what it delegated. See §12.4 and §12.5.

Some designs obtain bounded revocation freshness differently, by requiring the verifier to fetch a short-lived status assertion from the issuer before honouring a token. HDP does not, because that makes every verification depend on the issuer being reachable. The trade is deliberate: HDP keeps verification offline and leaves the freshness of revocation state to whoever populates it.

Revocation is per token, not per hop. There is no mechanism to revoke authorization for a single delegate in the middle of an otherwise valid chain while leaving the token valid; the token is revoked and, if the task is to continue, re-authorized (§6). Deployments that require per-delegate revocation SHOULD layer a capability system that supports cascade revocation at the application layer. Retaining accountability for each delegate, for example through distinct per-hop identifiers (§9.1), is what makes such application-layer revocation actionable.

A consequence of any revocation mechanism is that proof an action was authorized is not, by itself, proof that the authorization was still current when the action was taken. An auditor reading a token after the fact cannot tell from the token whether it had been revoked at a verifier; that information lives in the verifier’s state and SHOULD be retained alongside stored tokens where audit requires it.

The scope.max_hops field (§3.3) creates an incentive that works against the purpose of this protocol, and issuers need to understand it before setting the field. When an application gates actions on a valid chain and the hop budget is exhausted, an agent that still needs to delegate has two options: seek re-authorization (§6), or delegate without appending a hop. The second is off-record delegation. It produces a chain that verifies, an action that occurred, and no record connecting them, which is the worst outcome a provenance protocol can produce. A limit meant to constrain delegation instead constrains the recording of it.

The usual motivation for limiting delegation depth is to bound the cost of verifying, storing, or reasoning about long chains. That is a verifier concern and belongs at the verifier: a verifier MAY reject or flag chains longer than a locally configured limit, as a heuristic it controls and can adjust without reissuing tokens. Placing the limit in the token binds every verifier to a number chosen at issuance and hands the incentive above to every agent that carries the token.

Accordingly, issuers SHOULD omit max_hops unless the delegation budget is itself part of what the human declared and recording it has evidentiary value. Where the field is set, applications SHOULD make re-authorization readily available to agents that exhaust it, so that the honest path is not more costly than the off-record one. The field is retained in v0.1 because it is optional and because, where a human did declare a budget, the declaration is provenance.

An agent may hold more than one valid token whose scope covers the same resource: for example, one issued for Alice authorizing a read of a dataset and one issued for Carol authorizing an update to it. Applications MUST bind an action or attempted action to the task and token that actually triggered it, and agents MUST record it under that context. They MUST NOT select a different token merely because its scope would make the action appear authorized. A deviation from the triggering token’s scope SHOULD be recorded under that token, explicitly identified as an attempted, blocked, or observed violation in action_summary. Recording the deviation does not amend scope or assert that the principal approved it. If the triggering context is unknown, the application MUST preserve that uncertainty in its audit record rather than assign an unrelated principal.

HDP cannot detect a violation of this rule. A hop appended to the wrong token verifies at every step of the pipeline, because the pipeline establishes that the hop was recorded, not that it was recorded in the right place. The consequence is borne at audit: an update performed in Carol’s task but recorded under Alice’s token misattributes the triggering context and leaves Carol’s token silent. Conversely, an update that actually occurred in Alice’s read-only task belongs in Alice’s task record as a violation; it MUST NOT be moved to Carol’s token to make it appear permitted. A signed record alone cannot prove correct task attribution, and inclusion of an action MUST NOT be interpreted as proof that the principal approved it.

Where more than one task could legitimately initiate an action, selecting the initiating task is application policy. Applications SHOULD define and record that choice before execution, together with a request or event identifier. Once selected, the triggering context governs provenance even if the action deviates from its scope. Multi-principal delegation (§7) does not address this case: it covers tokens linked by parent_token_id that share a session_id, and the tokens here are unrelated. The neighbouring question of how a service decides which of several grants applies to a request is an authorization-layer question and is outside HDP (§1.1).

Prompt injection attacks attempt to cause an agent to act as if it received instructions from a legitimate principal, when in fact the instructions originate from adversarial content in the agent’s environment, such as a malicious web page or document. HDP mitigates but does not fully prevent this attack.

Applications SHOULD detect and block actions that contradict the human’s declared scope using their own enforcement mechanisms. The semantic comparison is application-defined. Blocking an action MUST NOT require suppressing its evidence: an HDP-aware agent SHOULD record the attempted action, the detected scope deviation, and whether it was blocked in action_summary under the triggering task’s token. If a violation is observed after execution, it SHOULD likewise be recorded as an observed violation, without claiming that the principal approved it or that the signature proves execution. The hop timestamp remains the extension time, not a backdated event time. If the token cannot be extended, for example because its hop budget is exhausted, the application SHOULD retain an integrity-protected incident record linked to the token digest and triggering request. HDP v0.1 adds no status field for this purpose; the distinction is explicit in the declaration.

The mitigation HDP provides is evidentiary: an HDP-aware agent records each delegation action it takes as a signed hop, so an action carried out under a legitimately issued token leaves an auditable record, supporting post-hoc detection of prompt injection. This mitigation depends on agents actually recording their actions; an agent that omits a hop is discussed in §10.4.

The security of all HDP guarantees depends on the confidentiality of the issuer’s Ed25519 private key. Implementations MUST:

HDP makes a strong architectural guarantee: a correct implementation of the seven-step verification pipeline requires no network calls, no registry lookups, and no third-party contact. The complete trust state required for verification is:

This guarantee is a property of the verification procedure, not of the single-key signing model of v0.1. Per-agent hop signing on the pattern described in §4.2 would preserve it, since the verifier would still resolve only the issuer’s key out of band.

This guarantee enables HDP verification in air-gapped environments, edge deployments with intermittent connectivity, and latency-sensitive contexts where a network round-trip before every action is unacceptable.

This document requests registration of the following HTTP header fields in the Hypertext Transfer Protocol (HTTP) Field Name Registry.

This document requests registration of the application/hdp-token+json media type in the Media Types registry, following the procedures of RFC 6838.

This document requests registration of the following entry in the Well-Known URIs registry, per RFC 8615.

The Intent Provenance Protocol (IPP) and HDP address the same root problem with different architectural trade-offs. The key differences are:

These are design choices, not defects. Deployments with reliable connectivity to a revocation service, existing DID infrastructure, and a requirement for revocation across token ancestry may prefer IPP. Deployments that prioritize offline operability, self-sovereignty, and minimal infrastructure may prefer HDP.

OAuth 2.0 Token Exchange defines a mechanism for exchanging one security token for another, including delegation and impersonation use cases. HDP and RFC 8693 are complementary rather than competing: RFC 8693 governs access token issuance and delegation in an OAuth 2.0 authorization server context, while HDP governs the provenance record that travels with an agentic task regardless of the authentication mechanism used.

HDP tokens do not replace OAuth access tokens. An agent framework MAY use OAuth 2.0 for resource authorization and HDP for delegation provenance simultaneously.

JSON Web Token provides a general-purpose signed claims format. HDP differs from JWT in three respects:

UCAN (User Controlled Authorization Networks) defines a capability-based authorization token system with chained delegation. HDP and UCAN share the concept of delegation chains but differ significantly in scope: UCAN is a general capability authorization system, while HDP is specifically a provenance record for human-authorized agentic tasks. HDP makes no claims about capability enforcement; UCAN tokens carry executable capabilities that are enforced by receiving systems.

A UCAN delegation records the authorization provenance of a capability: who delegated what to whom. UCAN’s separate Invocation and Receipt objects can record individual invocations and their results; HDP instead keeps the execution record inline in the delegation chain itself, as the signed action_summarydeclared at each hop, so that the human authorization and the subsequent declared actions travel together in a single offline-verifiable record. In this sense HDP complements capability systems rather than competing with them: a deployment MAY use UCAN, or ZCAP-LD, for capability delegation and HDP alongside it for the tamper-evident execution record.

ZCAP-LD expresses delegated authorization capabilities as Linked Data, with invocation and delegation rooted in a controller’s key. As with UCAN, a ZCAP-LD delegation chain captures the authorization provenance of a capability but not a record of the delegate’s subsequent actions. HDP neither defines nor enforces capabilities; it records the human authorization event and the subsequent execution history. Deployments that already use ZCAP-LD MAY use HDP alongside it to supply the execution audit trail ZCAP-LD does not itself provide.

The Open Digital Rights Language (ODRL) (W3C ODRL) is a W3C Recommendation for expressing permissions, prohibitions, and constraints. Several fields in HDP’s scope object (§3.3) overlap with concepts ODRL already defines: authorized_tools and authorized_resources correspond to ODRL actions and targets, network_egress and persistence map to ODRL permissions or prohibitions, and quantitative limits such as max_hops map to ODRL constraints.

HDP v0.1 deliberately retains a small, self-contained scope object rather than embedding an ODRL policy. The trade-off is explicit: the minimal object keeps tokens compact and implementable with only JSON and Ed25519, at the cost of the vocabulary reuse, policy composability, and tooling interoperability that ODRL provides. Deployments that already reason over ODRL policies will require a separate mapping to interpret HDP scopes.

A further limitation of the v0.1 scope object is that it is fixed at issuance. HDP has no attenuation: a delegate cannot narrow the scope at its own hop, because hops record actions rather than grants and a hop record has no field in which a narrower scope could be expressed. A delegate that wishes to pass on less than it received must obtain a new token from the issuer with a narrower scope. Per-hop caveats, which only the verifier and any attenuating agent would need to interpret, are planned for a future version.

Because the chain-of-custody mechanism is payload-agnostic (§1.5), a future HDP profile MAY carry an ODRL policy as its payload in place of the native scope object. Such a profile would gain a natural binding to the Verifiable Credentials Data Model 2.0 (W3C VC Data Model 2.0), whose termsOfUse property can carry ODRL policies. This binding is identified as future work and is not specified in this document.

A complete token with a two-hop delegation chain. Signature values are truncated for readability. This example validates against the published JSON Schema. The scope lists a resource for each tool that acts on one, and each hop names the resource it declares acting on; §3.3 explains why v0.1 cannot bind tools to resources structurally.

This document tracks the revisions of the Internet-Draft it is cited by. The token wire format has not changed across any revision listed here and remains HDP v0.1: a token issued under revision -00 has the same structure and the same signature payloads as one issued today. What has changed is what verifiers, implementers, and auditors are required to do with it. Revisions are listed newest first.