Tool Restraint

SkillAI & models

This skill adds a principle of restraint: the more tools an AI carries, the more ways it can fail and the harder every choice becomes. Once added, your AI can keep its tool set lean, skip tools it does not need, and stay dependable when you are connecting new tools. Reach for it when wiring tools or when your AI seems to be underperforming.

Available today. Use it from your connected AI after setup.

Add it, then apply its guidance whenever you are wiring tools into your AI or when the AI underperforms. Keep the tool set as small as the work allows.

Then ask your AI: use the Tool Restraint skill

What your AI can do with it

  • Keep an agent's tool set limited to what the task actually needs
  • Avoid adding tools that increase the chances of failure or confusion
  • Guide which tools to wire in when giving the AI new abilities
  • Spot when too many tools are behind an agent's underperformance
  • Reduce the decision load on the AI so it picks the right tool more often

What this skill tells your AI

The instructions your AI receives, as published by archive228/loopkit in skills/tool-restraint/SKILL.md and read by ahel’s review.

Loading an agent with 14 MCP servers "just in case" makes it slower and dumber, not more capable — tool-use agents hit sharp capability cliffs as cognitive load rises.

  • Enable only the servers the current work actually uses. Remove the rest.
  • Prefer official servers for credentialed tools; never install five speculatively.
  • Each tool's description eats context on every turn — fewer, sharper tools beat a junk drawer.
  • Before adding a write-scoped server, add a hook that logs every call. If the agent picks the wrong tool or thrashes, the fix is usually fewer tools with clearer descriptions, not a smarter model.

Signals

GitHub stars
754
Forks
126
Last commit
Jul 2026
Advanced
Catalog kind
skill
Gateway key
tool-restraint
Source
github.com/archive228/loopkit