https://github.com/shihchengwei-lab/workslop-checker
A tool for checking whether AI-written documents are ready to hand off, ship, publish, or use for decisions. Not an AI detector.
https://github.com/shihchengwei-lab/workslop-checker
ai chatgpt-custom-gpt claude-code document-quality document-review workslop
Last synced: 19 days ago
JSON representation
A tool for checking whether AI-written documents are ready to hand off, ship, publish, or use for decisions. Not an AI detector.
- Host: GitHub
- URL: https://github.com/shihchengwei-lab/workslop-checker
- Owner: shihchengwei-lab
- License: other
- Created: 2026-05-19T14:20:10.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2026-05-19T15:08:50.000Z (about 1 month ago)
- Last Synced: 2026-05-19T17:53:02.759Z (about 1 month ago)
- Topics: ai, chatgpt-custom-gpt, claude-code, document-quality, document-review, workslop
- Homepage:
- Size: 37.1 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# verify-doc / Workslop Checker
English | [ηΉι«δΈζ](README.zh-TW.md)
> A tool for checking whether an AI-written document is ready to hand off, ship, publish, or use for decisions.
> Not about detecting AI. About stopping AI output from dumping cleanup work onto whoever has to use it next.
---
## Try it directly (recommended)
No setup needed. Click the link:
**π https://chatgpt.com/g/g-6a0c5412dca48191be1632fac7523af4-verify-doc-workslop**
The GPT is bilingual and automatically matches the language of your conversation.
---
## What this solves
**Workslop** is a term coined by BetterUp Labs and Stanford Social Media Lab in their [September 2025 Harvard Business Review article](https://hbr.org/2025/09/ai-generated-workslop-is-destroying-productivity), defined as: "AI-generated work content that masquerades as good work but lacks the substance to meaningfully advance a given task."
Their research found:
- **40%** of employees received workslop in the past month
- **~2 hours** average to correct each piece
- **$186/employee/month** ($9M/year for a 10,000-person org)
This tool does not ask **"is this AI-written?"** β that question is hard to answer and often does not solve the real problem.
It asks **"will this document push cleanup work onto whoever has to use it?"** Can you hand it off, ship it, decide on it β or do you need to redo half the work first?
---
## How to use
### Most users
Click the GPT link above. Paste your document into the conversation. Ask whether it is ready to use.
### Claude Code users (developers)
Copy `claude-skill/en/SKILL.md` into your `~/.claude/skills/verify-doc/SKILL.md`. Restart Claude Code. Trigger with "verify this doc", "doc review", or "workslop check".
(Traditional Chinese version: [`claude-skill/zh-TW/SKILL.md`](claude-skill/zh-TW/SKILL.md).)
### Want to fork or customize?
See the README files in these folders:
- [Claude skill (English)](claude-skill/en/SKILL.md) / [Claude skill (zh-TW)](claude-skill/zh-TW/SKILL.md) β full ruleset
- [ChatGPT Custom GPT setup (English)](chatgpt-custom-gpt/en/README.md) / [ChatGPT Custom GPT setup (zh-TW)](chatgpt-custom-gpt/zh-TW/README.md) β Description / Conversation starters / Instructions, ready to paste into Custom GPT Editor
---
## Good for
- AI-written reports, SOPs, lesson plans, proposals, slide text, announcements, meeting notes, handoff documents
- Anyone deciding whether a document can be handed off, shipped, or used as a basis for decisions
## Not for
- Conversation quality checks (rating each chat response)
- Code review
- Chat logs, scratch notes, rough drafts, or personal notes not meant to be delivered
---
## Structure
```
workslop-checker/
βββ README.md # This file (English)
βββ README.zh-TW.md # Traditional Chinese
βββ LICENSE # CC-BY-4.0
βββ claude-skill/
β βββ en/SKILL.md # Full Claude Code skill (English)
β βββ zh-TW/SKILL.md # Full Claude Code skill (Traditional Chinese)
βββ chatgpt-custom-gpt/
β βββ en/README.md # Custom GPT setup (English) β Description/Starters/Instructions
β βββ zh-TW/README.md # Custom GPT setup (Traditional Chinese)
βββ docs/
βββ design-rationale.md # Why this design
βββ workslop-background.md # Workslop origin & context
```
---
## Feedback
Open a GitHub issue if the tool gets something wrong:
- **False positive**: a good document was flagged Red or Yellow
- **False negative**: workslop was rated Green
Please include:
- Document characteristics (type, scale, source β no need to paste the whole thing)
- The tool's verdict
- What you expected
---
## License
[CC-BY-4.0](LICENSE). Commercial use allowed. Attribution required.
---
## Credits
- **Workslop concept**: BetterUp Labs + Stanford Social Media Lab ([original HBR article, Sep 2025](https://hbr.org/2025/09/ai-generated-workslop-is-destroying-productivity))
- **Design**: [@shihchengwei-lab](https://github.com/shihchengwei-lab) in collaboration with Claude / Codex