context-bridge-dispatch

What problem does it solve?

This skill originally addressed the inefficiency of redundant context passing and uncoordinated output when running multiple AI coding agents in parallel, which caused wasted tokens, conflicting workstreams, and reduced output quality for complex development tasks.

Core Features & Use Cases

• Context Card Injection: Automatically provided relevant context to parallel agents, using a lightweight standard mode for 2 or fewer agents and a full Registry mode for 3 or more agents to avoid context bloat.
• Coordination Overhead Budget: Limited redundant context sharing between agents to reduce token waste and prevent overlapping work.
• Focus Mode: Kept parallel agents aligned on task boundaries to minimize off-topic output and rework.
• Use Case: Development teams running multiple parallel coding agents on large, interconnected codebases could use this skill to ensure each agent had only the context it needed, cutting down on wasted compute and inconsistent outputs.

Quick Start

Use the /agent-composer skill for all parallel dispatch with context injection, as this deprecated skill has been merged into agent-composer Step 3-a and is no longer actively maintained.