Preview unavailable. View source
Headroom: shrink what your agent reads, save tokens
Coding Agent Skills
Squeezes tool output, logs, files and long chats before they reach the model, so Claude Code, Codex and other agents spend fewer tokens with the same answers.
Install command
pip install "headroom-ai[all]" headroom wrap claude
Agent prompt
Use the "Headroom: shrink what your agent reads, save tokens" skill for my current task. GitHub source: https://github.com/headroomlabs-ai/headroom Fetch this source and read its README, applicable SKILL.md files, and referenced instructions before using it. Add the relevant skill to your own supported skills directory or use the documented installation method. Check the source and permissions before running any setup commands. Published installation command: pip install "headroom-ai[all]" headroom wrap claude Apply the skill to my current project: Squeezes tool output, logs, files and long chats before they reach the model, so Claude Code, Codex and other agents spend fewer tokens with the same answers. Follow the repository instructions, make the required changes, and verify the result. If my intended task is missing, ask me to describe it first.
Agents: Claude Code, Codex, Cursor