agent-organ-system

name	agent-organ-system
version	2
description	Use this skill when the Agent needs to map a request or failure to one of its long-term capability organs: executive control, proactive orchestration, self-correction, growth cycle, capability map, distillation metabolism, verification, or repository maintenance. It prevents dumping everything into memory and keeps each organ's responsibility clear.
allowed-tools	read_file list_directory

Agent Organ System

Purpose

Provide a stable organ map for the Agent's long-term capability architecture. Use it before adding or changing self-governance assets so each change lands in the right organ.

Organ Map

• Prefrontal cortex / executive control: agent-executive-control

  • Decides direct execution, read-only inspection, confirmation, or clarification.
  • Handles “继续 / 执行 / 落地 / 接管 / 放开手脚”.

• Motor coordination / proactive orchestration: agent-proactive-orchestration

  • Keeps work moving through task classification, closure, and handoff.

• Reflex correction / behavior repair: agent-self-correction

  • Handles user corrections, wrong boundaries, over-caution, missed persona, or repeated behavioral mistakes.

• Metabolism / distillation: self-distillation-metabolism

  • Turns experience into the correct asset layer: memory, skill, script, job, runtime, docs/archive.

• Growth cycle / learning loop: agent-growth-cycle

  • Decides whether an experience deserves durable change and verifies that growth landed.

• Skeleton / capability map: agent-capability-map

  • Determines where assets should live and avoids directory sprawl.

• Immune system / safety boundaries: SAFETY_BOUNDARIES.md

  • Protects credentials, destructive actions, downloads, restarts, source files, and high-impact operations.

• Reality check / verification: verification-before-completion

  • Prevents unsupported completion claims.

• External memory sync / repository maintenance: moviepilot-agent-git-maintenance

  • Keeps capability assets portable, auditable, and recoverable.

• Skill architecture / rule system governance: skill-architecture-governance

  • Designs, reviews, splits, merges, and prunes skills/rule assets using skill-based architecture principles.

Routing Rule

When a self-improvement request arrives:

1. Use agent-executive-control if the user is pushing execution or complaining about hesitation.
2. Use agent-self-correction if the user says the Agent misunderstood or behaved wrongly.
3. Use self-distillation-metabolism if the user asks to absorb, distill, burn away noise, or implement into all layers.
4. Use agent-growth-cycle if the task is about building future learning behavior.
5. Use agent-capability-map if the question is mainly “where should this live?”
6. Use skill-architecture-governance if the task is about good skill writing, skill-based architecture, routing source, thin shells, splitting/merging skills, or rule lifecycle maintenance.
7. Use verification-before-completion before claiming the organ change is complete.

Anti-Patterns

Avoid:

• treating memory as the only brain
• adding a new organ for a problem already covered by an existing skill
• mixing safety, persona, media workflow, and self-governance into one file
• creating diagrams or metaphors that do not change routing or execution