Quick Fire

Tired of your AI agent turning every answer into an essay? Put it on a word budget. Quick Fire limits every response to 10 words—or any number you choose—so your agent cuts the waffle and gets straight to the point. Set it once; it stays active for the conversation.

Install

Install the standard Agent Skill globally:

npx skills add thellmwhisperer/quick-fire -g

Target a particular supported agent when needed:

npx skills add thellmwhisperer/quick-fire -g -a codex
npx skills add thellmwhisperer/quick-fire -g -a claude-code
npx skills add thellmwhisperer/quick-fire -g -a pi

Hermes can install the skill directly:

hermes skills install thellmwhisperer/quick-fire/skills/quick-fire

Claude Code users can add the Quick Fire marketplace and install its plugin:

/plugin marketplace add thellmwhisperer/quick-fire
/plugin install quick-fire@quick-fire

Claude namespaces marketplace skills, so invoke it as /quick-fire:quick-fire 10 or stop it with /quick-fire:quick-fire stop.

Invoke

quick-fire       # 10-word session budget
quick-fire 60    # 60-word session budget
quick-fire stop  # return to normal, unrestricted responses

Invocation syntax depends on the agent. Codex normally exposes $quick-fire; Claude Code and Hermes expose /quick-fire. Pi and OpenCode users who require the exact /quick-fire command can install the corresponding file under integrations/ in addition to the standard skill.

Repository layout

• skills/quick-fire/: portable Agent Skill.
• integrations/pi/: Pi prompt template for exact /quick-fire invocation.
• integrations/opencode/: OpenCode command for exact /quick-fire invocation.

License

MIT © 2026 Javier Mellado.