A user has reported that their educational cybersecurity tool catalog, hosted on Hugging Face Spaces, was automatically flagged as abusive and paused. The space, named pentester-run-place, serves as a curated directory for open-source security research tools rather than containing malicious code.
The automated flagging likely resulted from security-related terminology found in translated README content included in a Python seed file during a recent deployment. The user has since removed the auto-run of that file to prevent resource spikes and is requesting a manual review to restore access.
This incident highlights how automated content filters can mistakenly identify educational or reference materials as abusive due to specific technical keywords.