Academy Glossary

Academy reference

Glossary for verifiable authority in agentic systems.

Use this page as a companion reference while moving through the Academy. It translates the recurring Mandaitor, identity, authorization, evidence, and governance terms into practical language for learners, reviewers, and builders.

Return to Academy overview Read Core Concepts

The Academy introduces ideas from identity standards, verifiable credentials, runtime authorization, protocol-based tool use, and AI risk management. The W3C DID Core specification defines decentralized identifiers as URI-based identifiers whose associated DID documents can express verification methods and services.1 The W3C Verifiable Credentials Data Model describes a model for tamper-evident claims exchanged between issuers, holders, and verifiers.2 The Model Context Protocol provides a common way for applications to connect language models to external tools and context, which makes explicit authorization boundaries increasingly important.3 NIST's AI Risk Management Framework frames AI governance as a socio-technical risk discipline rather than a purely technical checklist.4

This glossary keeps those background ideas close to Mandaitor's product vocabulary. It is intentionally shorter than an API reference: each entry explains what the term means, why it appears in the Academy, and where to go next for deeper reading.

Use case	How to use the glossary
First-time learner	Keep this page open while reading the beginner and identity lessons, especially when terms such as mandate, delegate, verifier, and credential appear together.
Builder or architect	Use the runtime and evidence entries to align mental models before reading implementation guides for mandate creation, verification, and agent registry integration.
Reviewer or governance stakeholder	Use the evidence and governance entries to connect dashboard signals, evidence packs, audit events, and escalation paths to real control objectives.

Authority and delegation

Authority

The bounded right to perform an action on behalf of another party. In Mandaitor, authority should be explicit, inspectable, and limited to a defined action context.

Academy use: Authority is the central question behind the Academy: may this agent's request be carried out, before anything reaches a tool or business system?
Related terms: Mandate, Principal, Delegate, Scope

Mandate

A digital, verifiable record of delegated authority. It connects who grants authority, who receives it, what action is allowed, and which constraints apply.

Academy use: Mandates are the Academy's main object for turning informal permission into a runtime-verifiable authority boundary.
Related terms: Principal, Delegate, Scope, Constraints, Proof-of-Mandate

Principal

The person, organization, service, or legal entity that owns the authority and grants it to another actor.

Academy use: The principal is the source of delegation. A mandate has no business meaning unless the principal is understood and trusted.
Related terms: Delegate, Subject, Mandate

Delegate

The actor that receives authority from the principal. In agentic systems this is often an AI agent, service account, workflow, or automation layer.

Academy use: The delegate is the actor whose action must be checked before execution, especially when it can call tools, APIs, or operational systems.
Related terms: Principal, Agent, Mandate

Scope

The exact action and resource boundary of a mandate. Scope answers what the delegate may do and where that permission applies.

Academy use: Scope keeps authorization narrow enough for runtime verification and later review instead of relying on broad platform access.
Related terms: Action, Resource, Taxonomy, Constraints

Constraints

Additional limits that control when and how a mandate is valid, such as time windows, financial limits, rate limits, approval rules, or escalation thresholds.

Academy use: Constraints explain why a credential or mandate may be technically valid but still not acceptable for a specific runtime action.
Related terms: Scope, Lifecycle state, Human escalation

Taxonomy

A structured vocabulary for domain-specific actions and resources. It gives mandate scopes a shared language instead of free-form permission text.

Academy use: Taxonomies help teams define reusable action names, resource patterns, and policy boundaries for a specific industry or workflow.
Related terms: Scope, Action, Resource

Identity and proof

Decentralized Identifier (DID)

A standards-based identifier that can be associated with verification material without depending on a single centralized identity provider.

Academy use: DIDs help the Academy explain how an agent, organization, or service can have a stable identity that can be resolved and verified.
Related terms: DID document, Verification method, Subject

DID document

The document associated with a DID. It commonly contains verification methods, service endpoints, and metadata needed to interact with or verify the DID subject.

Academy use: The DID document bridges an identifier and the verification material needed to decide whether a signature or credential should be accepted.
Related terms: DID, Verification method, Trust chain

Verifiable Credential

A tamper-evident digital statement made by an issuer about a subject. A verifier can check the credential's integrity and issuer context.

Academy use: Verifiable credentials explain how claims about mandates, authority, identity, or status can become portable proof artifacts.
Related terms: Issuer, Holder, Verifier, Claim, Presentation

Claim

A specific statement inside a credential, such as an agent identity, allowed action, mandate identifier, status, or issuance timestamp.

Academy use: Claims are the data points reviewers and systems inspect when deciding whether a proof says what it appears to say.
Related terms: Credential, Presentation, Selective disclosure

Presentation

A package of credential material shown to a verifier. A presentation may reveal all claims or only selected claims depending on the disclosure model.

Academy use: Presentations connect credential theory to the practical question of what proof a reviewer, partner, or system receives.
Related terms: Credential, Verifier, Selective disclosure

Trust chain

The sequence of issuers, keys, registries, policies, and business decisions that explain why a verifier should accept a proof.

Academy use: Trust chains help learners separate cryptographic validity from the broader decision that a proof is appropriate for a business context.
Related terms: Issuer, Verifier, DID document, Credential status

Runtime authorization and evidence

Runtime verification

The act of checking a proposed action at the moment an agent or service wants to perform it, rather than only reviewing permissions beforehand or logs afterward.

Academy use: Runtime verification is the control point where mandate scope, constraints, lifecycle state, and request context are evaluated before execution. In Mandaitor's execution model, the service that decides is also the one that acts.
Related terms: Verification decision, Execution, Mandate, Constraints

Verification decision

The result of a runtime authorization check, such as allowing, denying, or requiring escalation for a proposed action.

Academy use: Verification decisions teach learners that authorization should produce a reasoned outcome, not merely a silent technical pass or failure.
Related terms: Runtime verification, Reason code, Evidence pack

Proof-of-Mandate

A signed statement that a valid mandate authorized one specific operation, issued when Mandaitor sends the write.

Academy use: Proof-of-Mandate is the bridge between runtime authorization and later auditability, because it preserves why a specific action was allowed. The execution receipt then shows what it did.
Related terms: Mandate, Execution receipt, Verifiable Credential, Evidence pack

Execution

One operation an agent asked Mandaitor to perform, and what became of it. Mandaitor reads the record, decides on what it read, writes once, and reads the result back.

Academy use: Executions move the action itself out of the agent: the agent can only ask, so the mandate is the only way to the target system.
Related terms: Tool, Execution receipt, Approval, Mandate

Execution receipt

Mandaitor's signed statement of what an operation did: the record before and after, who approved it, and the Proof-of-Mandate it was made on.

Academy use: The receipt completes the evidence: the Proof-of-Mandate says the write was allowed, the receipt says what happened.
Related terms: Proof-of-Mandate, Execution, Evidence pack

Tool

One typed business operation in one system, such as releasing a purchase requisition in SAP. A connector holds Mandaitor's credential for the system; the agent never sees it.

Academy use: Tools make the boundary concrete: an agent can ask only for operations that have one, and each declares the facts it reads and the fields it returns.
Related terms: Execution, Taxonomy, Scope

Request-bound decision

A single-use authorization for exactly one request, which an integrator's enforcement point redeems immediately before its own system acts.

Academy use: Request-bound decisions keep the same boundary when Mandaitor cannot perform the operation itself: the authorization never passes through the agent.
Related terms: Verification decision, Execution, Mandate

Audit event

A recorded event that captures something important in the mandate or verification lifecycle, such as creation, status change, verification, denial, or evidence generation.

Academy use: Audit events are the raw timeline that evidence packs, dashboards, and governance reviews rely on.
Related terms: Audit trail, Evidence pack, Compliance dashboard

Evidence pack

A review package that organizes proofs, audit events, decisions, reason codes, and context so a human or downstream system can understand what happened.

Academy use: Evidence packs turn runtime authorization into reviewable operational evidence for security, compliance, customer support, and dispute resolution.
Related terms: Proof-of-Mandate, Audit event, Compliance dashboard

Selective disclosure

A proof pattern where only selected claims are revealed while hidden claims remain protected without breaking verification.

Academy use: Selective disclosure helps explain how Mandaitor can support auditability while reducing unnecessary exposure of sensitive identifiers or tenant data.
Related terms: Presentation, Credential, Proof-of-Mandate

Governance and adoption

Compliance dashboard

A product surface that summarizes mandate status, verification outcomes, evidence availability, and review signals for teams responsible for oversight.

Academy use: The dashboard lesson teaches that summarized status must always connect back to proof, reason codes, and underlying evidence.
Related terms: Evidence pack, Audit event, Governance control

Governance control

A policy, procedure, system check, reviewer action, or escalation path that reduces operational, security, legal, or compliance risk.

Academy use: Governance controls connect Mandaitor's technical evidence to risk owners who need accountable decisions, not just logs.
Related terms: Human escalation, Evidence pack, AI risk

Human escalation

A control path where a request is paused, routed, or reviewed by a person because policy, risk, confidence, or business context requires human judgment.

Academy use: Human escalation reminds learners that verifiable automation should support accountable operations rather than remove all human responsibility.
Related terms: Constraints, Governance control, Verification decision

Lifecycle state

The current status of a mandate or proof-relevant object, such as active, pending approval, suspended, revoked, or expired.

Academy use: Lifecycle state matters because an otherwise well-formed mandate should not authorize action when it is no longer active for that context.
Related terms: Mandate, Constraints, Compliance dashboard

Product adoption path

A staged route for moving from conceptual understanding to a bounded Mandaitor evaluation, integration plan, operational review, and broader rollout decision.

Academy use: The adoption path helps prospects and implementation teams turn Academy concepts into a concrete first use case and evaluation checklist.
Related terms: Development status, Governance control, Evidence pack

How this glossary relates to the rest of the docs

The glossary is a learning aid, not a substitute for implementation guidance. When you need step-by-step product instructions, move from the term to the corresponding concept page, guide, or API reference. When you need business interpretation, move from the term back into the Academy lesson that teaches the decision context.