learn-from-examples
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external repositories and local paths, which serves as a significant ingestion point for indirect prompt injection. Malicious instructions embedded in source files (e.g., READMEs or code comments) could attempt to influence the agent's behavior during the synthesis and output generation phases. The skill lacks explicit instructions for the agent to treat analyzed content as data rather than instructions. Ingestion points: Git URLs and local paths defined in Step 1. Capability inventory: Network access for git cloning, file reading, and writing to the local filesystem in Step 7. Sanitization: The instructions do not mention sanitizing or escaping the snippets extracted from the repositories.
- [EXTERNAL_DOWNLOADS]: The skill performs git clone operations to fetch external data from remote URLs provided by the user.
- [COMMAND_EXECUTION]: The workflow involves executing system commands such as git and potentially codegraph to survey and search the repository structure.
Audit Metadata