[BUG] Content filtering blocks extraction of peer-reviewed medical genetics paper (false positive)
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
What I was doing
I am conducting a systematic review on AI/ML for rare genetic disease diagnosis at Vanderbilt University Medical Center. I use Claude Code to extract and analyze scientific papers from PDFs.
The blocked operation was writing a markdown text extraction of the following peer-reviewed paper:
- Title: "Artificial Intelligence in the Genetic Diagnosis of Rare Disease"
- DOI: 10.1016/j.cll.2022.09.023
- PMID: 36764805
- Publisher: Elsevier
Actual behavior
The API returns HTTP 400 with "Output blocked by content filtering policy", preventing the Write tool from executing. The entire extraction workflow is halted.
What Should Happen?
Expected behavior
Writing a faithful text extraction of a peer-reviewed medical review article should not trigger content filtering. This paper discusses computational tools for genetic diagnosis — it has no connection to harmful content categories.
This is a review article published in a mainstream medical journal. It surveys NLP tools for phenotype extraction (MetaMap, cTAKES, ClinPhen) and phenotype-driven variant prioritization tools (Exomiser, Xrare, DeepPVP). It contains zero harmful content — it is a clinical informatics literature review.
Error Messages/Logs
### Environment
- **Claude Code version:** Latest (Feb 2026)
- **Model:** claude-opus-4-6
- **OS:** Windows 11 Home 10.0.26200
- **API Request ID:** `req_011CY1QerWLrjhdBVXsrPbUS`
### Error
API Error: 400
{
"type": "error",
"error": {
"type": "invalid_request_error",
"message": "Output blocked by content filtering policy"
},
"request_id": "req_011CY1QerWLrjhdBVXsrPbUS"
}
Steps to Reproduce
- Read a medical genetics PDF using Claude Code's Read tool (visual/image mode) — succeeds
- Ask Claude to write the extracted text content to a markdown file using the Write tool — blocked by content filter
The Read tool successfully rendered all 15 content pages as images. Claude processed the content correctly. The block occurred only when attempting to output the extracted text as a markdown file.
Claude Model
Opus
Is this a regression?
Yes, this worked in a previous version
Last Working Version
any version before this
Claude Code Version
2.1.39 (Claude Code)
Platform
Anthropic API
Operating System
Windows
Terminal/Shell
Windows Terminal
Additional Information
_No response_
This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗