daily-papers-fetch
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python scripts (
fetch_and_score.pyandenrich_papers.py) using shell commands. It incorporates user-defined parameters (such as the number of days to fetch) into the script arguments. - [INDIRECT_PROMPT_INJECTION]: The skill fetches and processes data from external, untrusted sources including the arXiv API, HuggingFace API, and external HTML/PDF documents.
- Ingestion points: Paper metadata, abstracts, and full-text content are retrieved from arXiv, HuggingFace, and general web URLs (SKILL.md).
- Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following instructions potentially embedded in the fetched paper content.
- Capability inventory: The skill executes local scripts via
python3and writes structured JSON data to the/tmpdirectory. - Sanitization: The instructions do not describe any sanitization, filtering, or validation of the external content before it is parsed and used to enrich the data.
Audit Metadata