Glossary

Term	Definition
Agent	Autonomous computational entity with identity, capabilities, and an execution loop that produces decisions. See Concepts.
Identity	Cryptographic (e.g. Ed25519) keypair; fingerprint is the stable agent id. Used to sign and verify messages. See Concepts.
Message	Communication primitive: sender, optional recipient, topics, payload, optional signature. See Concepts.
Topic	Semantic namespace for routing, pool scope, and discovery. See Concepts.
Task	Unit of work with objective, constraints, and lifecycle (pending → assigned → completed/failed/cancelled). See Concepts.
Pool	Scoped sub-network with membership, admission, and optional governance. See Concepts.
Decision	Action emitted by an agent (e.g. SendMessage, ClaimTask, ReportTask). Executor runs them. See API/core.
Executor	Runs decisions (e.g. StandardExecutor). See API/runtime.
Discovery	Service and queries to find agents by topic/capability. See Concepts.
Governance	Model for resolving disputes in a pool (e.g. democratic, dictatorial). See Concepts.
Admission	Policy that controls who can join a pool (open, whitelist, token). See API/policy.
Trust	Per-agent scores used for discovery threshold and context. See Concepts.
Transport	Layer for sending and receiving messages (local, TCP, WebSocket). See Design.
Store	Key-value persistence for pools, tasks, discovery. See Store backends.
Capability	Declared skill or constraint of an agent for discovery. See Concepts.