Glossary | Palveron Docs

Term	Definition
Agent	An AI system that performs tasks autonomously or semi-autonomously (chatbot, coding assistant, automation)
Attestation	Cryptographic anchoring of a trace hash on the Flare blockchain
BYOK (planned)	Bring Your Own Key — customer-managed encryption keys for trace storage. On the post-launch roadmap; not yet generally available.
BYOW	Bring Your Own Wallet — customer-managed blockchain wallet for attestations
Capability Model	Permission model defining per agent which actions (tools) are allowed, denied, or require approval
Dispatcher	Module that sends governance events to external systems (Slack, Teams, ServiceNow, Jira)
Evidence Package	Exportable proof bundle with policies, traces, blockchain proofs, and incident reports
Flare	Blockchain network used for anchoring governance evidence
FRIA	Fundamental Rights Impact Assessment — per Art. 27 EU AI Act
Governance Request	A single verify call to the platform = 1 trace = 1 billing unit
Guardrail	A natural-language policy that inspects and acts on LLM traffic
HITL	Human-in-the-Loop — requiring human approval as part of the governance process
Ingestion Path	How traffic enters the platform (Gateway, SDK, MCP, Webhook, Browser Guard)
Integrity Hash	SHA-256 hash of trace content — basis for tamper detection
Merkle Proof	Cryptographic proof that a hash is part of a Merkle tree anchored on the blockchain
Neural Instruction	The natural-language directive that defines a policy's detection logic
PII	Personally Identifiable Information — data that can identify an individual
Responsibility Chain	Three documented responsible persons per agent: Liability, Technical, Approval
Trace	A single request-response pair processed by the platform, including decision and metadata
VerifyRequest	The universal governance contract — every ingestion path normalizes to this format