marketing-science-writing
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions direct users to clone a repository from an unverified GitHub account (liyuanbo1024/marketing-science-writing.git) to install the skill.
- [INDIRECT_PROMPT_INJECTION]: The skill's workflow processes untrusted user-supplied content (research topics, variable descriptions, and data file contents) to generate code and LaTeX manuscripts, creating a vulnerability surface.
- Ingestion points: Research topics provided via trigger phrases and external data loaded via the DATA_PATH environment variable.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to treat user-provided data as potentially adversarial or to ignore embedded instructions.
- Capability inventory: The skill enables the agent to read files (pd.read_csv), write code (Python/R), and generate LaTeX manuscripts as described in SKILL.md.
- Sanitization: The skill does not mention any sanitization or validation of the research topics or data values before they are interpolated into code or documentation.
Audit Metadata