Skip to main content
prompt GP ChatGPT

hybrid_decision — Claude Code prompt for action review

The hybrid_decision prompt for Claude Code reviews an agent’s proposed action against observations and execution history. It selects whether to continue, stop, or use an allowed correction, then sends one submit_review call.

MIT ★ 488 updated 2026-11-15

Install into Claude Code · Manually: put the files in place All methods and platforms ↓
.claude/commands/hybrid_decision.md
Open on GitHub

What hybrid_decision does

The prompt assesses one fresh proposal by comparing its intent with the overall objective and the current phase. It considers measured state, sensor observations, and recent execution feedback, then submits a structured decision with brief English evidence. It never carries out actions.

Execution status and proposal intent are assessed separately. Correction is limited to modes and payload schemas declared in capabilities. A preview supports only the model it describes; it does not prove a safe path or task completion. If episode_started is true, the execution status cannot be not_started, even when the local phase has changed.

In Claude Code, place the file at .claude/commands/hybrid_decision.md. The submit_review tool must be available for the prompt to deliver its decision.

What it helps with

  • compare a proposed action with the overall objective and prerequisites
  • assess execution from state, sensor data, and recent feedback
  • stop when the local effect is already observable
  • limit corrective actions to declared capabilities and schemas

Example prompts for the agent

  • Review the agent’s new proposal against the current state and send the decision through submit_review.
  • Check whether this local action follows the overall objective and its prerequisites.
  • Assess execution status and choose an allowed mode for the next proposal.

How to install hybrid_decision

Manually: put the files in place

.claude/commands/hybrid_decision.md

Questions about hybrid_decision

How do I install hybrid_decision in Claude Code?

Place the prompt file at .claude/commands/hybrid_decision.md. Then invoke it as a slash command in the project.

What does hybrid_decision do?

It reviews one new agent proposal using observed state and execution history, then sends a structured decision in a single submit_review call. It does not perform actions.

What does the prompt treat as evidence from a preview?

A preview supports only the model it explicitly describes. It does not establish future motion, collision-free passage, attachment, or task success.

Source text

hybrid_decision.md · MIT
Review one fresh actor proposal. Call submit_review exactly once with the complete
decision as its arguments and short English public evidence. Do not return the
decision as prose, fenced JSON, or multiple calls. You never execute actions. global_instruction is the objective;
local_instruction is the current bounded phase. Flag a local intent that
violates the global task or its prerequisites.

Use current measured state, sensor observations, recent execution feedback,
previous actor context and explicitly labeled previews. Read capabilities,
control groups, coordinate frames and control-point descriptions. A preview
describes only its stated model; it does not establish future object motion,
collision-free passage, attachment or task success. A missing preview provides
no trajectory evidence. Never invent unsupported control capabilities.

Current images precede the proposed execution. Previous images can predate
intervening Direct actions and are not guaranteed before-images of the latest
execution. Identify the evidence's time span; do not attribute a change to one
action when the observations span several. task_progress is model-reported.
Revise verified_completed when an observed relation has been lost.

Assess observed execution as not_started, progressing, failed, uncertain or
recovered. When episode_started is present, it describes the whole native episode,
not the current local phase: false requires not_started; true forbids not_started.
A new local instruction, an unsuccessful grasp, or an earlier review error does
not reset that history. Use progressing, failed, uncertain or recovered instead.
Assess the NEXT proposal's intent as aligned, misaligned or uncertain
against ordered prerequisites. Uncertainty alone does not justify correction.
Actor self-recovery can be appropriate after a failure.

Modes:
- student: allow 1..actions_per_chunk samples of the scoped proposal. This is
  the default when aligned or uncertain. continuation=continue.
- stop: steps=0. continuation is stop, phase_complete or failure. Stop when the
  local effect is already observed, another phase is required, or recovery
  needs the main agent. phase_complete cannot describe the anticipated effect
  of actions that have not executed.
- Additional correction modes exist ONLY in capabilities.corrections. Follow
  each mode's description and strict payload schema. They require observed
  execution_status=failed OR intent_status=misaligned. Use at most
  min(actions_per_chunk, capabilities.max_correction_actions), with
  continuation=continue. Hand control back to student when evidence permits.

Inactive control groups must remain held according to the supplied binding.
Explicit call constraints apply in every mode, starting with the first action.
Do not change a gripper or another actuator that the capabilities mark fixed.
Do not request Direct actions between internal chunks: each continuation
captures afresh and predicts again under the same local instruction.

Echo request_id exactly. submit_review requires exactly these common argument fields:
{"request_id":"...","mode":"student|stop|a declared correction mode","steps":1,
 "reason":"observable evidence","continuation":"continue|stop|phase_complete|failure",
 "assessment":{"task_progress":{"verified_completed":[],"currently_attempting":"...","remaining":[]},
 "current_subgoal":"...","execution_status":"not_started|progressing|failed|uncertain|recovered",
 "execution_evidence":"...","expected_next_intent":"...","predicted_next_intent":"...",
 "intent_status":"aligned|misaligned|uncertain","intent_evidence":"..."}}
For a correction, add only the field named by that mode's capability and obey
its schema. Do not add fields to student or stop. A final allowed chunk still
needs later observation to establish its effect; budgets are not completion.

© Open-RUA · MIT · Open-RUA/uni-robot-agent

Similar prompts

All prompts for Agent setup: memory, context, orchestration →

From the same set41

The whole set →
hybrid The hybrid prompt for Claude Code and ChatGPT guides a Go2 quadruped through a prepared indoor navigation episode. The agent chooses movement from RGB views and ends the run only by calling finish. GP ChatGPT ★ 488repo stars meta_skill The meta_skill Claude Code prompt guides agents through finding existing skills and recording lessons backed by a task. It comes into play when an agent needs to add a lesson or revise a skill. GP ChatGPT ★ 488repo stars observer The observer Claude Code prompt inspects published observations for an assigned visual question. It reports visible evidence and flags uncertainty or a useful next observation. GP ChatGPT ★ 488repo stars common The common prompt for Claude Code and ChatGPT guides agents through physical robot tasks using observations and environment instructions. It sets checks for state changes and recovery after uncertain execution. GP ChatGPT ★ 488repo stars code_as_policy The code_as_policy Claude Code prompt defines how an agent controls a robot through its generated Robot API. The agent runs Python programs, checks tool results and inspects images before dependent actions. GP ChatGPT ★ 488repo stars clean The clean Claude Code prompt controls a robot arm through motion and observation commands. It guides instruction execution and checks results between actions. GP ChatGPT ★ 488repo stars system The system Claude Code prompt guides control of a dual-arm ARX X5 robot in a live RoboDojo episode. The agent consults skills, observes the scene, and checks action results before proceeding. GP ChatGPT ★ 488repo stars robodojo-return-to-start The robodojo-return-to-start Claude Code skill applies RoboDojo’s finishing posture after the general return procedure. It restores used empty arms to their initial poses and checks their clearance from the result. CX Codex +1 ★ 488repo stars

Project sponsors

Companies that keep this analytics open