How does this skill handle file editing safety?

The skill integrates with File Safety locks to ensure that multiple agents do not attempt to edit the same files simultaneously, preventing conflicts.

How does the skill decide which agent to use?

It evaluates rules defined in .synapse/delegate.md based on keywords in your request and the specific file types involved, applying the first matching rule.

What happens if a task doesn't match any delegation rules?

The skill applies a fallback behavior which can be configured to ask the user for direction, default to a specific agent, or require manual handling.

What is the difference between Orchestrator and Passthrough mode?

In Orchestrator mode, Claude coordinates the task, waits for the delegated agent's response, and integrates it. In Passthrough mode, the task is directly forwarded to the target agent.

Can I manually delegate a task to a specific agent?

Yes, you can use the @agent pattern (e.g., @codex or @gemini) to manually route tasks regardless of the automatic delegation rules.

Multi-Agent Task Delegation

Name: Multi-Agent Task Delegation
Author: s-hiraoku

bys-hiraoku

生产力与工作流

Automates task routing and coordination between multiple AI agents using natural language rules and specialized responsibilities.

关于

This skill enables Claude to function as an orchestrator within the Synapse A2A ecosystem, intelligently delegating specific tasks to specialized agents like Codex or Gemini. By defining natural language rules and keyword matches in configuration files, users can create sophisticated workflows where Claude analyzes requirements, routes coding tasks to specialized models, triggers research-focused agents, and integrates the final results. It features robust priority management, error handling, and File Safety integration to ensure reliable and safe cross-agent collaboration.

主要功能

Agent status verification and priority-level management (1-5)
Interactive TUI configuration for managing agent responsibilities
0 GitHub stars
Multiple operational modes including Orchestrator, Passthrough, and Manual
Integrated File Safety protocols for multi-agent file editing
Rule-based task routing using keyword and file pattern matching

使用场景

Automating documentation research by delegating web-search tasks to specialized agents during development
Coordinating large-scale project updates across multiple agents with synchronized response handling
Routing complex refactoring tasks to a dedicated coding agent while Claude handles architectural review

关于

主要功能

Agent status verification and priority-level management (1-5)
Interactive TUI configuration for managing agent responsibilities
0 GitHub stars
Multiple operational modes including Orchestrator, Passthrough, and Manual
Integrated File Safety protocols for multi-agent file editing
Rule-based task routing using keyword and file pattern matching

使用场景

Automating documentation research by delegating web-search tasks to specialized agents during development
Coordinating large-scale project updates across multiple agents with synchronized response handling
Routing complex refactoring tasks to a dedicated coding agent while Claude handles architectural review