stop-that-shit
Sets a clear scope for your AI, cuts extra work and filler, and can add a guard that blocks out-of-scope actions.
- Ecosystem
- Agent Skills
- Skill category
- Coding
- Creator / organization
- lennney(许可证署名 Stop That Shit contributors)
What it is
stop-that-shit is a toolkit by lennney, licensed to Stop That Shit contributors, currently at release 0.2.5. It has three parts: a Skill that limits task scope, an optional execution Guard, and STSS for tighter answers. You agree with it on the job through review, answer, monitor or change; with the Guard installed, it checks each tool call against those limits and blocks anything outside them. The author's goal is to keep the AI on the current task and drop pointless extra work and defensive filler. It supports Codex, Claude Code, OpenCode, Hermes Agent CLI, Pi and Oh My Pi.
Who it's for:Developers tired of AI making changes nobody asked for and padding its answers.
Before you start: Follow INSTALL.md to install the Skill alone or the plugin with the Guard. The full Guard needs a registered hook that you explicitly trust, requires certain host versions, and changes what tool calls are allowed to do. The Skill alone doesn't block anything. Docs are mainly in Chinese. License: MIT.
Based on the project's public documentation. Not tested by AI Inspiration World.