pre-work-research
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill utilize standard project exploration commands such as
ls,find,cat, andgrep. These are used to locate and read documentation and code patterns within the repository, which is standard behavior for development-focused agents. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository files and external web resources during the research process, creating a surface for indirect prompt injection.
- Ingestion points: Local repository files read via
catandgrep, and external content retrieved viaWebFetchandWebSearchtools. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the research protocol.
- Capability inventory: The skill possesses capabilities to read files, search/fetch web content, and post comments to GitHub issues using the
ghCLI. - Sanitization: There are no explicit sanitization or validation steps for the content retrieved before it is summarized or used to inform the implementation approach.
Audit Metadata