FeedHive/ai-slop-detector

选用指南

AI 根据 README 整理 · 2026/10/14 19:46:05

目标用户

面向使用AI写作工具的用户，以及希望检查文本是否带有AI痕迹的写作者。

查看 README 依据

# AI Slop Detector 120 patterns that make writing read like AI slop, in 9 categories. Each pattern has a short explanation, a bad example and a plain rewrite. Give the list to your AI before it writes, or check a draft against it afterwards. To check a draft in your browser, paste it into the free [AI Slop Detector](https://www.feedhive.com/tools/ai-slop-detector).

使用方式

可通过浏览器在线工具粘贴文本检查，或作为Agent技能安装到Claude Code等工具中使用，也可将README粘贴到聊天应用让AI避免这些模式。

查看 README 依据

# AI Slop Detector 120 patterns that make writing read like AI slop, in 9 categories. Each pattern has a short explanation, a bad example and a plain rewrite. Give the list to your AI before it writes, or check a draft against it afterwards. To check a draft in your browser, paste it into the free [AI Slop Detector](https://www.feedhive.com/tools/ai-slop-detector).

The list doesn't try to tell whether a person or an AI wrote a text. People write these patterns too, and a model can be told to avoid them. A single pattern is normal in human writing; several in one paragraph is what makes a text read as slop. ## Use it with your AI **Claude Code, Codex and other agents.** Give the agent this repository and ask it to install the skill: ```text Install the AI Slop Detector skill from https://github.com/FeedHive/ai-slop-detector ``` Or install it yourself for Claude Code: ```sh git clone https://github.com/FeedHive/ai-slop-detector.git

cp -r ai-slop-detector/skills/ai-slop-detector ~/.claude/skills/ ``` **ChatGPT, Claude and other chat apps.** Paste this README into the chat and ask it to avoid every pattern in what it writes next. To check a draft, paste the README and the draft and ask for every pattern it finds, with the line it found it in. **Your own tools.** [`patterns.json`](patterns.json) has the full list as structured data. Edit it and run `node scripts/build.mjs` to regenerate this README and the skill. ## Categories | Category | Patterns | |---|---| | [Vocabulary](#vocabulary) | 20 |

部署要求

若作为技能使用，需要Claude Code或Codex等支持技能的Agent环境；若自行安装技能，需克隆仓库并复制到~/.claude/skills/目录。

查看 README 依据

The list doesn't try to tell whether a person or an AI wrote a text. People write these patterns too, and a model can be told to avoid them. A single pattern is normal in human writing; several in one paragraph is what makes a text read as slop. ## Use it with your AI **Claude Code, Codex and other agents.** Give the agent this repository and ask it to install the skill: ```text Install the AI Slop Detector skill from https://github.com/FeedHive/ai-slop-detector ``` Or install it yourself for Claude Code: ```sh git clone https://github.com/FeedHive/ai-slop-detector.git

cp -r ai-slop-detector/skills/ai-slop-detector ~/.claude/skills/ ``` **ChatGPT, Claude and other chat apps.** Paste this README into the chat and ask it to avoid every pattern in what it writes next. To check a draft, paste the README and the draft and ask for every pattern it finds, with the line it found it in. **Your own tools.** [`patterns.json`](patterns.json) has the full list as structured data. Edit it and run `node scripts/build.mjs` to regenerate this README and the skill. ## Categories | Category | Patterns | |---|---| | [Vocabulary](#vocabulary) | 20 |

已知限制

该清单不判断文本是否由AI或人类撰写，因为人类也可能使用这些模式；单个模式在人类写作中正常，多个模式集中出现才可能被视为AI痕迹。

查看 README 依据

The list doesn't try to tell whether a person or an AI wrote a text. People write these patterns too, and a model can be told to avoid them. A single pattern is normal in human writing; several in one paragraph is what makes a text read as slop. ## Use it with your AI **Claude Code, Codex and other agents.** Give the agent this repository and ask it to install the skill: ```text Install the AI Slop Detector skill from https://github.com/FeedHive/ai-slop-detector ``` Or install it yourself for Claude Code: ```sh git clone https://github.com/FeedHive/ai-slop-detector.git