本文へ移動
cccskills
無料GitHub で公開

tool-restraint

Don't over-equip an agent with tools/MCP servers. More tools = more ways to fail and a higher cognitive load. Use when wiring tools or an agent underperforms.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md937 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Tool Restraint

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.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

a11y-pass

無料

Catch the accessibility failures that ship in almost every AI-built UI. Use after building any interactive component.

日本語の概要は準備中です。原文の説明を表示しています。

Archive228/loopkit7552026年7月15日 更新

Before compaction Loopkit extracts decisions into claude-decisions.json (machine-readable). Read it alongside claude-progress.txt at session start — prose is for humans, JSON is for the loop.

日本語の概要は準備中です。原文の説明を表示しています。

Archive228/loopkit7552026年7月15日 更新

Review a diff against the goal spec assuming the code is BROKEN. The reviewer that lives in the maker's head always agrees with itself — this pulls review into a hostile, separate pass. Invoke after every code change before marking work done.

日本語の概要は準備中です。原文の説明を表示しています。

Archive228/loopkit7552026年7月15日 更新

Verify that an endpoint checks ownership, not just authentication. Use on any handler that reads or mutates user data.

日本語の概要は準備中です。原文の説明を表示しています。

Archive228/loopkit7552026年7月15日 更新

Find the exact commit that introduced a bug. Use when something worked before and broke, and you don't know which change did it.

日本語の概要は準備中です。原文の説明を表示しています。

Archive228/loopkit7552026年7月15日 更新

Before picking new work, smoke-test the last "completed" feature. If it's broken, revert and re-open it before touching anything else. Kills the "looks shipped, isn't shipped" bug across sessions.

日本語の概要は準備中です。原文の説明を表示しています。

Archive228/loopkit7552026年7月15日 更新

Archive228 のスキルをすべて見る

このスキルの問題を報告する