Stats
Actions
Tags
Open the interactive AgentDispatch configurator to add/remove models, set tier spoofing, scan Ollama, and manage providers. Use when user says "configure dispatch", "add model", "dispatch config", "dispatch setup", or "/dispatch-configure".
How this skill is triggered — by the user, by Claude, or both
Slash command
/claude-agent-dispatch:dispatch-configureThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Interactive configuration for AgentDispatch models, providers, and tier spoofing.
Interactive configuration for AgentDispatch models, providers, and tier spoofing.
This is an interactive terminal tool — it requires direct user input. Tell the user to run it themselves.
First, find the plugin directory. Check DISPATCH_PLUGIN_DIR env var, or look for agent-dispatch/scripts/configure.py relative to the project root.
Then tell the user to type this in the prompt (the ! prefix runs it interactively in the session):
! python <plugin-dir>/scripts/configure.py
For user-level config (applies to all projects):
! python <plugin-dir>/scripts/configure.py --user
To find the path, use:
Bash: echo $DISPATCH_PLUGIN_DIR
Or check .mcp.json for the dispatch server's DISPATCH_PLUGIN_DIR env value.
The configurator provides a menu-driven interface:
ollama list, pick by numberDo NOT try to run configure.py via Bash tool — it needs interactive stdin from the user. Always instruct the user to run it with ! prefix.
npx claudepluginhub markx95/claude-agent-dispatch --plugin claude-agent-dispatchCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.