marketing-pipeline-automation
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by processing data from untrusted external sources.
- Ingestion points: The
lib/scraper/news-crawler.tsmodule fetches content from TechCrunch, a16z, and potentially other sources based on a user-provided keyword. - Boundary markers: The skill lacks explicit delimiters or boundary markers (e.g., XML tags or clear instructions to ignore embedded commands) when interpolating
researchDatainto the prompt viaJSON.stringify(researchData)inlib/ai/content-generator.ts. - Capability inventory: The system possesses network access (calling OpenAI/Anthropic APIs) and file system access (writing rendered videos to the
public/videosdirectory). - Sanitization: There is no evidence of sanitization or filtering applied to the crawled news content before it is processed by the LLM.
- [EXTERNAL_DOWNLOADS]: The installation instructions direct users to clone an external repository (
https://github.com/pennydinh/marketing-pineline-share.git) and install its dependencies vianpmoryarn. This is standard for project setup but introduces third-party code into the execution environment.
Audit Metadata