SAFR whitepaperp. 8
Action Trace
The steps an agent actually took to reach a proposal: tool calls, data retrieved, checks performed.
Glossary
Every term, labelled by whether it comes from the SAFR whitepaper or from OpenSAFR. Page references point to the whitepaper, version 1.0 (July 2026).
SAFR whitepaperp. 9
Agent Identity
The component that binds a proposed action to a verified, registered agent before any other evaluation.
SAFR whitepaperp. 9
Agent Registry
The authoritative record of registered agents against which identity is verified; may be internal or network-level.
SAFR whitepaperpp. 11–12
Audit Log
A tamper-evident, append-only record of every governance decision.
SAFR whitepaperp. 11
Auto-Execute
Outcome: the action proceeds without human intervention.
SAFR whitepaperp. 8
Context Metadata
Agent identity, applicable mandate, account or system state, and policy constraints attached to an envelope.
SAFR whitepaperpp. 9–10
Controls Repository
The institution's configurable rulebook of controls drawn from policy, regulation, product rules, and mandates.
SAFR whitepaperp. 10
Deny
Outcome: the action is rejected before execution, with a specific reason recorded.
SAFR whitepaperpp. 10–11
Disposition Engine
The component that evaluates an action against applicable controls and resolves it to one of four outcomes.
SAFR whitepaperp. 11
Escalate
Outcome: the action is held pending human review.
SAFR whitepaperp. 16
Evidence Quality
A control category setting minimum confidence and evidence requirements for autonomous execution.
SAFR whitepaperp. 8
Governance Envelope
The structured record an agent submits with each proposed action: action, action trace, and context metadata.
SAFR whitepaperp. 10
Mandate
A machine-readable delegation of authority from a user to an agent, defining what it may do and within what limits.
SAFR whitepaperp. 11
Observe
Outcome: the action proceeds but is flagged and logged for later review.
SAFR whitepaperp. 4
SAFR
Safeguards for Agentic Finance at Runtime: an industry reference approach for governing agent actions before they execute.