version: "1.0.2" name: agent-planning-loop description: Review unsettled plans with independent scouts when the user requests multi-agent planning or alignment. disable-model-invocation: true

Agent Planning Loop

Run a bounded planning debate when requested. Return one judged recommendation; do not implement, approve the plan, or let scouts ask the user questions. Preserve the full intended scope of a supplied plan unless the user selects a subset or a material decision blocks honest full-plan execution.

Set a brief charter: goal, non-goals, context, permitted worker actions, and stop condition. Use scouts when preparing worker prompts and judging their evidence. Main-session judgment owns the recommendation.

Default to at most two rounds: discovery/challenge, then a narrowed follow-up only if it can change the recommendation. Stop sooner when a plan has clear acceptance and proof, the remaining items are implementation defaults, or a user decision blocks progress. A deeper autonomous loop requires the user's request.

Output

Return a review packet, not a transcript.

Include: planning loop status, what ran, rejected/parked items, one recommended plan with scope/non-goals/acceptance/verification/risk, alternatives only when material, user decisions, Plan artifact:, Recommended next action:, and Shortcut:.

For plan-file loops, the Recommended next action should normally execute or preflight the whole plan through the appropriate route. It may name the first phase as the starting point, but should not make that phase the whole approved scope unless the user asked for that narrowing or a stop gate blocks the rest.

If no material plan remains, say so directly and explain what evidence caused the stop. Do not fill the packet with weak alternatives. Still include Recommended next action and Shortcut; use park/none when there is genuinely no useful next step.

Approval Handling

If the user approves the review packet, do not rerun the planning loop unless their approval changes scope. Route to the single Recommended next action. Treat short replies such as LGTM, approve, sounds good, or do it as approval for that action and preserve the packet's plan artifact, scope, verification, user decisions, parked items, and stop condition.

If the user pushes back on part of the recommendation, treat that as new charter input. Rerun only the affected scout stage, not the whole loop.