Authority Grant
The explicit, bounded, revocable and recorded instrument by which a principal permits a named agent to commit specified classes of action, up to stated limits, for a stated term.
- StatusProposed by QIS (not yet industry usage)
- AdoptionExperimental
- LayerGovernance
Why it matters
Every autonomous action is an exercise of delegated authority, and at present the grant is almost never a distinct artifact — it is implied by a tool manifest, an IAM policy, or nothing at all. Naming it separately is the precondition for governing it. The historical analogue is a letter of marque: defined scope, defined term, accountability running back to the grantor.
This term is a QIS Canonical Extension. It does not currently appear in the Model Context Protocol specification, the A2A (Agent2Agent) specification, the AIEOG Shared AI Lexicon, or any instrument in our standards register. We are proposing it because we have documented the gap it fills, and we label it as ours everywhere it appears. Adopt it, argue with it, or ignore it — but do not cite it as industry consensus, because it is not.
Where it comes from
QIS Canonical Extension. Proposed to fill a gap: no reviewed instrument in the standards register defines the artifact that confers permission on an agent, and the AIEOG Lexicon contains no term for it.
Proposed here; instruments its origin note discusses
- AIEOG Shared AI LexiconUS Treasury · FBIIC (Financial and Banking Information Infrastructure Committee) · FSSCC (Financial Services Sector Coordinating Council)Official text In the standards register
What it is not
These are routinely confused with Authority Grant. The distinctions have consequences for how a system is governed.
- Tool Manifest
The manifest is capability. The grant is permission. An agent can reach a payment API and be permitted to move nothing.
- IAM policy
An IAM policy governs a credential's technical access. A grant governs an agent's commercial authority and names the principal.
- Guardrails
Guardrails are applied by the operator. A grant is issued by the principal and is the thing an examiner asks to see.
How it connects
What it needs
- Authority Grant requires Agent Identity Canon term
- Authority Grant is put into practice by Authority Envelope Canon term
- Authority Grant is limited by Fiduciary duty Canon term
- Authority Grant is completed by Governed Quantity Canon term
What relies on it
- Authority Grant sets limits on Agent Identity Canon term
- Authority Grant is a precondition for Agent Identity Canon term
- Authority Grant sets limits on Prompt Injection Canon term
- Authority Grant is a precondition for Agent Decision Record Canon term
- Authority Grant sets the authority or limits for Agent Canon term
MARQUE questions it addresses
- AuthorityWhat is the agent permitted to commit, and who granted that?
Not to be confused with
- Tool Manifest Canon term
Related framework objectives
Connected through the MARQUE (a reference framework organized around eight questions) crosswalk, an editorial mapping. These objectives from the Financial Services AI Risk Management Framework are assigned to the same questions that this term addresses. The ones sharing most of this term's vocabulary are shown first, as a reading aid. The assignment is to the question, not to the term: the source does not map this term to any objective directly.
Question: Authority · 39 control objectives assigned
- GV-2.1.1AI Risk Management Roles and ResponsibilitiesGovern
- GV-2.1.2AI Risk Communication and EscalationGovern
- MG-1.3.2Responsibility and Decision CriteriaManage
Cite and reuse
Cite the identifier QIS-TERM-00020, not the URL: identifiers are stable, URLs may change. This entry is free to read, quote and index under the dual license. Corrections sent to editor@qisfund.com are published.
Technical details: identifier, graph relationships and record
| Predicate | Target identifier | Target | Predicate meaning |
|---|---|---|---|
addresses | marque:authority | Authority | The subject speaks to the object as a question or concern. |
constrains | term:agent-identity | Agent Identity | The subject bounds what the object may do. |
notEquivalentTo | term:tool-manifest | Tool Manifest | The two are routinely conflated and are distinct. |
requires | term:agent-identity | Agent Identity | The subject imposes the object as a precondition. |
requiredBy | term:agent-identity | Agent Identity | Read the other way round: the object requires the subject. The subject imposes the object as a precondition. |
constrains | term:prompt-injection | Prompt Injection | The subject bounds what the object may do. |
requiredBy | term:decision-record | Agent Decision Record | Read the other way round: the object requires the subject. The subject imposes the object as a precondition. |
implementedBy | term:authority-envelope | Authority Envelope | The object is a concrete realization of the subject. |
relatedTo | term:consent-grant | Consent Grant | An association too weak or too general for a stronger verb. |
relatedTo | term:authority | Authority | An association too weak or too general for a stronger verb. |
relatedTo | family:construction | Construction | An association too weak or too general for a stronger verb. |
constrainedBy | term:fiduciary-duty | Fiduciary duty | Read the other way round: the object sets limits on the subject. The subject bounds what the object may do. |
completedBy | term:governed-quantity | Governed Quantity | Read the other way round: the object completes the subject. The subject is the form in which the object's deferred content (a schedule, a record) is actually produced; the object is not usable without it. |
governs | term:agent | Agent | Read the other way round: the object operates under the subject. The subject operates under the authority or constraint of the object. |
- Canonical identifier
QIS-TERM-00020- Tier
- QIS Canonical Extension
- Domain · layer
- Authority · Governance
- Semantic aliases
- None recorded
- First published
- 2026-08-02
- Last reviewed
- 2026-08-02 · 180-day cycle