Stats
Actions
Tags
From dispatch
This skill should be used when the user invokes "/dispatch:opus <task>", asks to "run this with opus", "use opus for this", "dispatch to opus", or needs deep reasoning for complex architecture, hard debugging, or critical code review. Dispatches the given task to an Opus model subagent and returns the result.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dispatch:opusThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Dispatch the task in `$ARGUMENTS` to an Opus model subagent.
Dispatch the task in $ARGUMENTS to an Opus model subagent.
Opus is the highest-capability, highest-cost tier. Reserve it for tasks where reasoning depth materially changes the outcome: architecture design, non-obvious root-cause debugging, critical security/integrity reviews, multi-constraint trade-off analysis.
$ARGUMENTS is empty, ask the user: "What task should I dispatch to Opus?"Agent({
model: "opus",
description: "<one-line summary of the task>",
prompt: "<the full task from $ARGUMENTS>"
})
If the task is routine (standard code edit, simple explanation, lookup), suggest
/dispatch:sonnet or /dispatch:haiku — Opus cost is only justified when Sonnet
would produce shallower results that matter. Don't refuse to run Opus if the user
explicitly chose it, but note the mismatch briefly.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub jobrien127/plugin-marketplace --plugin dispatch