The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Automated security scans confirmed the presence of malicious URLs within the skill's data infrastructure, specifically in the discovery and search results files.
Evidence: The domain ijels.com, referenced in _workspace/discovery/papers.json, is flagged as part of a botnet infrastructure.
Evidence: The URL https://ijels.com/upload_document/issue_files/17IJELS-10720257-NotBorn.pdf is explicitly identified as a malicious file.
Context: These links are provided as authoritative source material for the agent to reference, increasing the likelihood of the agent or user visiting these high-risk domains.
[PROMPT_INJECTION]: The core instructions in SKILL.md and AGENTS.md were analyzed for behavioral overrides, safety bypasses, and roleplay jailbreaks. No such patterns were detected. The skill instructions focus legitimately on the philosophical frameworks of Judith Butler.
[DATA_EXFILTRATION]: No scripts, commands (such as curl or wget), or instructions for accessing sensitive local file paths or environment variables were found.
[COMMAND_EXECUTION]: The skill does not contain any executable scripts or instructions for arbitrary shell command execution.

judith-butler