What problem does it solve?
When users @mention local CLI coding agents like Cursor, Codex, or OpenCode, vague one-line requests produce poor results and direct shell invocation bypasses the agent's reasoning. This Skill ensures every delegation is analyzed, expanded into a complete task, and routed through the local_agent tool.
Core Features & Use Cases
- Task Formulation: Expands short user requests into complete task descriptions with goals, context, acceptance criteria, and boundaries before delegation.
- Multi-Agent Orchestration: Detects installed agents with the status action, runs single delegations, or executes parallel delegations with run_many across OpenClaw, Hermes, Cursor, Codex, OpenCode, and Trae.
- Sequential and Parallel Workflows: Orders dependent tasks sequentially (e.g., implement then test) and runs independent tasks in parallel, then merges results.
- Use Case: A user types "@cursor implement the API @codex write integration tests". The Skill splits the request, formulates two complete tasks, runs them in dependency order via local_agent, and summarizes each agent's output.
Quick Start
Ask the agent to delegate a coding task by saying something like "@cursor refactor the login module in src/auth to use JWT sessions" and it will formulate and dispatch the task through local_agent.
