Bounded Agent Lifecycles
How Niyatna Agents receive scoped work, execute within boundaries, and report proof back.
A Niyatna Agent lifecycle begins when an intent is assigned and ends when the company receives proof and report-back.
Agents are useful because they can carry work forward, but they must remain bounded by role, context, permissions, and proof requirements.
Lifecycle phases
- Intent intake: The agent receives the goal, context, permissions, and acceptance criteria.
- Route selection: Niyatna chooses the right model, tool, data source, or human handoff.
- Execution: The agent performs the scoped work.
- Self-check: The agent gathers evidence and checks the work against the acceptance criteria.
- Review gate: Risky or external actions pause for approval.
- Report-back: The result, proof, and memory update return to the company system.
Bounded autonomy
Niyatna Agents can use powerful tools, but authority stays legible. The company should always know what the agent was asked to do, what it touched, and what proof it returned.