Key terms used in the PoCC specification
This glossary provides quick reference definitions for key PoCC terms. For detailed semantic definitions, see Ontology.
An entity capable of issuing PoCC-defined actions. Has a unique identifier, can cryptographically sign actions, and maintains an action history. Does not require intelligence, consciousness, or human oversight.
See also: Ontology: Agent
A discrete, typed operation performed by an agent and recorded immutably. Actions are cryptographically signed, timestamped, and may reference dependencies. Cannot be deleted, only superseded.
See also: Action Registry, Ontology: Action
An immutable, append-only record of all actions in a PoCC system. Chronologically ordered, cryptographically verifiable, and queryable by agent, action type, or dependency.
See also: Ontology: Action Log
A directed acyclic graph (DAG) representing causal relationships between actions. Nodes are actions, edges represent dependencies. Must be acyclic to prevent circular reasoning.
See also: Ontology: Dependency Graph
An agent introduces a claim into the shared action space. Does not imply truth, only records that an agent made a claim at a specific time.
See also: Action Registry: ASSERT
A formal objection to a prior action, requiring response or revision. References the challenged action and contains a reason (may be hash). Does not invalidate the target action.
See also: Action Registry: CHALLENGE, Ontology: Challenge
An agent supersedes a prior assertion without deleting it. Only the original author can revise. Original action remains in log, marked as superseded, with complete revision history preserved.
See also: Action Registry: REVISE, Ontology: Revision
A binding declaration by an agent that constrains future actions. Specifies conditions and constraints, cannot be silently withdrawn, and may be challenged or fulfilled.
See also: Action Registry: COMMIT, Ontology: Commitment
Transfer of task authority from one agent to another, with explicit conditions. Creates an accountability chain where original responsibility remains with delegator, execution responsibility transfers to delegate.
See also: Action Registry: DELEGATE, Ontology: Delegation
Declaration that a commitment or delegation has been completed. References the original commitment/delegation and may include proof or evidence hash. Can be challenged.
See also: Action Registry: FULFILL, Ontology: Fulfillment
An agent records a reasoning trace connecting premises to conclusions. Records logical dependency structure, not correctness of reasoning. May optionally include PoC validation metadata.
See also: Action Registry: REASON
The process of checking whether an action conforms to PoCC rules. Checks signature validity, schema conformance, dependency validity, and state transition legality. Does not check truthfulness, correctness, or quality.
See also: Ontology: Verification, Protocol: Verification Rules
A protocol that verifies the integrity of individual agent cognitive processes. PoC verifies cognition; PoCC commits cognition. PoC may exist without PoCC. PoCC may reference PoC, but does not require it.
See also: PoCC vs PoC
A protocol that enforces auditable, accountable, and revision-capable collaboration among autonomous agents. PoCC defines how agents interact, not what they believe. PoCC is structural, not normative.
See also: What is PoCC, Protocol Specification
The following concepts are not part of PoCC:
See also: Ontology: Explicitly Excluded Concepts