principalchain.com

The ordered sequence of human principals, organizational entities, and authorized AI systems that collectively authorize an AI agent's actions, from the root human authority through any intermediate delegators to the acting agent itself, forming the complete accountability lineage for any agent operation. The principal chain is the fundamental governance structure for attributing AI agent actions to responsible human principals. Every operation performed by an agent must be traceable through a valid principal chain to a recognized human or institutional root.

The formal enrollment of a human or institutional entity as the apex authority in one or more principal chains, establishing their identity credential, contact information, accountability scope, and the governance framework under which they exercise root authority. Root principal registration is performed by the organization's identity management system and requires verification of the principal's identity to a defined assurance level. Registered root principals are published to the mandate transparency register as publicly accountable parties.

A verifiable claim issued by the principal chain registry confirming the current depth of a specific principal chain — measured as the number of delegation links between the root principal and the acting agent — together with the depth limit encoded in the root delegation credential. Chain depth attestations are consumed by authorization systems that apply differential trust or capability restrictions based on delegation depth. Attestations are timestamped and expire after a short validity window to prevent replay against stale chain configurations.

A standardized sequence of cryptographic operations by which a relying party validates the integrity, completeness, and current validity of a principal chain presented by an acting agent before honoring the agent's authority claims. The protocol verifies each link's signature, confirms scope attenuation compliance across links, checks revocation status of all credentials, and validates depth limit compliance before issuing a chain verification result. Chain verification is a mandatory precondition for any operation requiring delegated authority above a defined risk threshold.

The governance principle establishing that each principal at a non-terminal position in a principal chain bears partial accountability for all actions taken by agents below them in the chain, proportional to the scope of authority they delegated, even when those actions were not directly instructed. Intermediate principals must implement monitoring proportional to the authority they delegate and cannot disclaim accountability by asserting ignorance of subordinate actions within their delegated scope. Accountability obligations are codified in the delegation credential and accepted upon issuance.

A protocol optimization technique that produces a compact proof of authority for presentation to relying parties by summarizing multiple delegation links into a single composite credential without sacrificing verifiability, reducing the overhead of chain transmission in performance-sensitive environments. Compacted chains reference the full chain's immutability anchor and include a digest of each omitted link, enabling any verifier to reconstruct and validate the full chain from the anchor and digest set. Compaction may only be applied to the middle links of a chain; the root and terminal credentials must always be presented in full.

The identification of delegation credentials whose issuing principal has been deregistered, incapacitated, or revoked, leaving the credential holder with authority that is no longer backed by a valid principal chain, creating an accountability gap that must be remediated through chain repair or credential revocation. Orphaned authority detection is performed by the principal chain registry at regular intervals and upon notification of principal status changes. Detected orphaned credentials are immediately flagged as invalid pending governance review.

Continuous automated surveillance of all active principal chains to detect revocations, expirations, scope violations, and depth limit breaches in real time, alerting the governance system before affected agents perform unauthorized operations based on stale chain state. Monitoring agents query the chain registry at defined intervals and subscribe to revocation event streams to ensure near-real-time awareness of chain validity changes. Integrity alerts include the specific chain identifier, the detected condition, and a recommended governance response.

A standardized rating assigned to a principal's identity credential reflecting the rigor of the identity proofing and authentication processes used to establish and maintain it, used to determine the maximum delegation scope and chain depth the principal may authorize. Higher assurance levels permit broader authority delegation and deeper chain construction, while lower assurance levels restrict delegation to low-risk, low-scope authority transfers. Assurance levels are verified during root principal registration and revalidated at defined intervals.

A formal authorization record enabling an agent operating under one principal chain to perform specific actions on behalf of a principal in a different chain, without merging the two chains or creating a subordination relationship between their respective root principals. Cross-chain bindings are time-limited, scope-restricted, and mutually signed by authorized representatives of both chains. They are recorded in both chains' audit trails and published to the principal chain registry for transparency.

The maximum set of capabilities an agent may exercise at a given position in a principal chain, computed as the intersection of the agent's intrinsic capabilities, the scope granted by its delegation credential, and the residual scope permitted by the depth-constrained attenuation of all preceding chain links. Capability ceilings are computed at binding time and enforced by the agent's policy enforcement layer as hard limits that cannot be exceeded regardless of runtime instruction content. Capability ceiling violations are security incidents.

A tamper-evident, chronologically ordered record of all significant events in a principal chain's lifecycle — including registration, delegation issuances, revocations, expirations, integrity alerts, and cross-chain bindings — maintained independently of the chain participants to provide an objective accountability record. The audit log is signed by the principal chain registry at each entry and anchored to an immutable ledger at regular intervals to prevent retroactive modification. Log retention periods are defined by governance policy and regulatory requirements.

The cryptographic signature applied by a delegating principal to the delegation credential they issue, binding their identity to the specific scope and terms of the delegation and preventing subsequent repudiation of the delegation event. The signature algorithm and key management requirements are specified by the organization's credential policy and must meet the assurance level of the delegating principal's identity credential. Delegation link signatures are verified during principal chain verification as evidence that each link was established by a legitimate, authorized party.

The real-time propagation of a principal revocation event to all downstream agents, relying systems, and registries that hold credentials or authority derived from the revoked principal, ensuring that the accountability gap created by revocation is immediately visible to all affected parties. Broadcast messages are signed by the revoking authority and include the revocation reason, effective timestamp, and a list of directly affected credentials. Systems receiving a revocation broadcast must immediately suspend operations dependent on the revoked authority pending governance resolution.

The governance design requirement that principal chains should be constructed with the fewest delegation links necessary to achieve the required authority transfer, minimizing accountability diffusion, verification overhead, and the attack surface created by intermediate delegation points. Applying this principle requires mandate authors to evaluate whether intermediate delegation links provide genuine governance value or merely reflect organizational convenience. Chain architectures that violate the minimal chain principle are flagged during mandature conflict pre-screen review as governance inefficiencies.