extend-domain-standards
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill researches external documentation and analyzes repository code, creating a surface for indirect prompt injection from potentially poisoned external sources.
- Ingestion points: Documentation is fetched via
WebFetchand repository files are read during the exploration phase. - Boundary markers: While no explicit string delimiters are defined, the process enforces logical separation through a contrast table.
- Capability inventory: The skill utilizes network read tools and file system write tools.
- Sanitization: Security is maintained through mandatory manual user review at the contrast, drafting, and writing stages.
- [SAFE]: The skill incorporates strong security best practices, including
disable-model-invocation: trueto prevent autonomous background activity and strict adherence to user-named scopes, preventing unsolicited file modifications.
Audit Metadata