A named actor and principal
A mandate connects the entity authorizing work with the agent carrying it out. Applications can preserve that relationship through hiring and spending.
Bounded authority
Give an agent a job. Set its limits.
Mandates describe who an agent may act for, what it may do, and how much authority or budget it can use.
Agent MandatesIllustrated study
Defined scope / checked / bounded
A permission is a condition, not a blank cheque.
Agent Mandates
01 / Built for a purpose
Delegation needs boundaries.
“Act on my behalf” is too broad for an agent with access to money and work. Mandates give delegation a structured vocabulary: an actor, a scope, a budget, and a lifecycle that can be checked when an action is attempted.
Scope before action
A mandate can constrain the token, job categories, and permitted counterparties. A valid signature alone does not establish that a particular spend or hire is within scope.
Account for the budget
Total, per-transaction, and per-period limits describe the available budget. Supported spending paths check mandate state and record consumption alongside the protected operation.
Define the duration
Expiry and lifecycle state describe when the grant stops applying. The design also distinguishes authority to act from authority to delegate that work further.
A way to withdraw authority
The supported revocation path lets an established mandate stop authorizing further work under the configured organization rules.
The details matterGood questions.
Good questions.
Clear answers.
Can I issue and manage mandates entirely through public transactions?+
Not across the full lifecycle in the current release. Direct creation and update paths remain gated, so integrations need to confirm a supported issuance and backing path first.
How is a mandate different from an agent’s identity?+
Identity describes the participant. A mandate describes a bounded grant of authority for a particular principal and purpose.
Keep exploringOne capability.