Skills
.Work. You
Rest

interaction-autonomy

Community

Control agent autonomy with five interaction modes.

Software Engineering#collaboration#interaction#checkpoint#autonomy#mode-switching#approval-gates#supervised
Authortylerjrbuell
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill lets teams control how much autonomy an agent has versus human oversight by providing five explicit interaction modes, runtime mode switching, and approval checkpointing to prevent unsafe or undesired actions.

Core Features & Use Cases

  • Five interaction modes: autonomous, supervised, collaborative, consultative, and interrogative to cover background automation, supervised deployments, live collaboration, advisory scenarios, and debugging.
  • Runtime mode management: enable .withInteraction(), switch modes via InteractionManager, and scope mode changes per agent ID.
  • Approval checkpoints & kill switch integration: pause/resume via checkpoint tools and kill switch for human approvals before irreversible actions (file writes, API calls).
  • Use case: Deploy a supervised agent that checkpoints plans before performing file writes so an operator can review and resume execution.

Quick Start

Enable the interaction layer on your agent and set the initial mode to supervised so the agent pauses for approval before irreversible actions.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: interaction-autonomy
Download link: https://github.com/tylerjrbuell/reactive-agents-ts/archive/main.zip#interaction-autonomy

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.