Agent Loop
An Agent Loop is a controlled cycle. The agent checks state, acts with an allowed tool, reads the result, then stops or continues.
Another example
A planner reads schema data, proposes one safe check, and records the result. It stops at its step limit without touching production.
Main parts
- 01State and goal
- 02Bounded action
- 03Observation and stop
Use it when
Use it when each new result changes the next action. Set rules for stopping, cost, permission, and failure.
Do not use it when
Avoid endless retries, vague success, unchecked tool effects, and work after authority or evidence runs out.
Name used in code
state → decide → tool → observe → stop/continue Before you ship
Check the model and prompt versions, source data, versioned evaluation set, and measures. Verify links to evidence, tool permissions, privacy, common failures, decline-to-answer and fallback behavior, human review, monitoring, cost, speed, and rollback.
Outcome: Use or evaluate Agent Loop to make AI behavior measurable and tied to evidence. User context: A planner reads schema data, proposes one safe check, and records the result. It stops at its step limit without touching production. AI method or concept: Agent Loop. Why it fits: Use it when each new result changes the next action. Set rules for stopping, cost, permission, and failure. Do not use it when: Avoid endless retries, vague success, unchecked tool effects, and work after authority or evidence runs out. AI requirements: Define the input and source evidence. Set model and tool permissions. Use versioned evaluation data and measures. Define failure, decline-to-answer, privacy, speed, and cost limits. Operational safeguards: Keep a clear trace and hide sensitive log data. Show users a safe fallback. Mark steps that need human review. Define how to roll back the model or prompt. Acceptance criteria: Record baseline and target measures on a versioned evaluation set. Test edge cases and hostile inputs. Verify fallback, monitoring, permissions, and rollback. Evidence and limits (evidence boundary): No single official standard controls the whole term. Its meaning may change across teams, platforms, or frameworks. Unknowns to confirm: Target task, model and version, evaluation owner, source data, risk limit, tool permissions, and production fallback.
Check this request
Review the current use of Agent Loop. Definition: An Agent Loop is a controlled cycle. The agent checks state, acts with an allowed tool, reads the result, then stops or continues. Release checks: Check the model and prompt versions, source data, versioned evaluation set, and measures. Verify links to evidence, tool permissions, privacy, common failures, decline-to-answer and fallback behavior, human review, monitoring, cost, speed, and rollback. Before changing code, report the evidence you found, gaps, severity, and the smallest safe fix.