From ravn-ai-toolkit
Launches a background sub-agent to execute tasks (via /parallel), allowing concurrent work. Returns results on completion.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ravn-ai-toolkit:parallelThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Run the given task in a background sub-agent so you can continue working on other things.
Run the given task in a background sub-agent so you can continue working on other things.
The user's input is:
$ARGUMENTS
Determine whether it is a skill invocation or a plain text task.
If the input starts with / (e.g., /promptify, /review-pr, /pr-comments-address TICKET-123), it is a skill invocation:
/) and any remaining text as the skill arguments./cc fix auth bug, tell the agent: "Use the Skill tool with skill 'cc' and args 'fix auth bug'."If the input does NOT start with /, pass it directly as the Agent tool prompt.
Use the Agent tool with run_in_background: true. Do NOT block on the agent's result. Continue responding to the user immediately after launching it. When the agent completes, report the result.
User: "/parallel /promptify audit all skills against our findings doc."
Expected behavior: Launch a background agent that invokes the promptify skill with args "audit all skills against our findings doc."
User: "Fix the bug in the checkout flow."
Expected behavior: Do not use parallel; execute the task directly in the foreground.
run_in_background: true is set and the prompt is a non-empty string./ prefix does not match any available skill./help.npx claudepluginhub ravnhq/ai-toolkitLaunches parallel subagents via the Task tool with run_in_background: true to concurrently implement features, analyze directories, or run independent tasks.
Parallelizes independent work (N slices, reviews, or tasks) across subagents using worktree isolation. Directs fan-out, scatter-gather, and comprehensive review workflows without agent confusion.
Dispatches parallel subagents for 2+ independent tasks with no shared state. Use when tasks are well-defined and have no sequential dependency.