generate-writeup
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
uv runcommand to execute local Python scripts for various processing steps including slide fetching, image conversion, and text extraction. These commands incorporate user-provided file paths and URLs as arguments.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as video transcripts and slide text to generate the final blog post write-up. - Ingestion points: Data is ingested from external YouTube transcripts (Step 3) and slide text extracted from PDF/URL sources (Step 5).
- Boundary markers: The generation prompt in Step 7 lacks explicit delimiters or instructions to treat variables like
TRANSCRIPTandSLIDE_TEXTas untrusted data. - Capability inventory: The skill can execute local scripts, read and write to the file system, and fetch content from external URLs.
- Sanitization: There is no evidence of sanitization or filtering applied to the transcript or slide text before it is interpolated into the LLM prompt.- [EXTERNAL_DOWNLOADS]: The skill fetches external content from user-specified URLs for both slides (PDF/RevealJS) and video transcripts (YouTube).
Audit Metadata