prune

仓库创建 2026年7月26日最近提交 13 小时前SkillHot 收录 11 小时前

这个 Skill 做什么

审计项目里的 agent 配置，把过时、没用、冲突或放错位置的 harness 文件找出来并建议清理，适合定期收敛长期堆积的上下文。默认只生成报告；加 --apply 时会把删除变成可审阅的提案，不会擅自删除。流程会先委托 /doctor 处理可推导的内容，并报告 bonsai 自身的上下文占用。

▸ 展开 SKILL.md 英文原文

Audit this project's agent configuration and remove what no longer earns its keep. Finds stale, unused, contradictory, and misplaced harness artifacts, delegates CLAUDE.md trimming to /doctor, and reports bonsai's own context footprint.

Stars

Forks

仓库内 Skill

积累中

7 日增星

安装 / 使用

给你的 Agent 一句话（通用）

帮我安装这个 skill：https://raw.githubusercontent.com/hartz89/bonsai/main/skills/prune/SKILL.md

或 curl 直取 SKILL.md

curl -fsSL "https://raw.githubusercontent.com/hartz89/bonsai/main/skills/prune/SKILL.md"

SKILL.MD 节选查看完整文件 ↗

# bonsai:prune

The half nobody else builds. A harness that only grows gets worse at month six — every resident token
spends a finite attention budget, so removal is as valuable as addition.

Default is report-only. `--apply` proposes removals through the normal review gate; it never deletes
unilaterally.

## Procedure

1. **Delegate first.** Invoke `/doctor`. It already trims derivable content from a checked-in
   `CLAUDE.md`, finds unused skills, MCP servers, and plugins, and flags slow hooks. Everything it
   covers is out of scope here — report what it found and move on.

2. **Scan deterministically.**

   ```
   python3 ${CLAUDE_PLUGIN_ROOT}/scripts/prune_scan.py --project . --plugin-ro

via SKILL·HOT · 数据来自 GitHub 公开信息 · 原文版权归作者所有