podcast-guest-research
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a local bash script
scripts/crawlora.shas a wrapper for API calls. The script implements several security best practices, including input validation for the API key, path whitelisting via regex, and preventing curl from accessing local files. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, creating a potential surface for indirect prompt injection attacks.
- Ingestion points: Podcast metadata and third-party web content retrieved from the Crawlora API, specifically through search and scrape endpoints (SKILL.md, reference/endpoints.md).
- Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious content within the fetched data (SKILL.md).
- Capability inventory: The agent can execute
scripts/crawlora.shto perform further network operations based on the data it processes (scripts/crawlora.sh). - Sanitization: There is no evidence of filtering or sanitization being applied to the text content retrieved from external sources before it is processed by the agent.
Audit Metadata