run

name	run
description	Execute a named agent role with optimized instructions. Explicit invocation only ($run). Never trigger implicitly.

Act as an orchestrator spawning a named subagent. Perform no preliminary actions — invoke immediately.

Input: $ARGUMENTS

Directives:

1. Parse: The first word of $ARGUMENTS names the subagent. Everything after the first word verbatim becomes the task instructions.
2. Optimize: Refine the task for token and context efficiency. Strip conversational fluff; use precise, direct language. If no task is provided, proceed with an empty prompt.
3. Execute: Spawn the $1 subagent with the optimized instructions. Wait for it to complete.
4. Present: Return the result directly, without preamble or commentary.
5. Critical: STOP. Perform no further operations after presenting results.

Rules:

• Present results as-is.
• Never modify results unless explicitly requested.