atlas-grill
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes a
SOURCE-MANIFEST.jsonfile that references external upstream repositories on GitHub for source tracking and integrity. This is a standard mechanism for managed skills to handle versioning and updates from their original authors. - [INDIRECT_PROMPT_INJECTION]: The skill identifies contradictions and edge cases by reading and analyzing local project files, such as source code and documentation. This creates a potential ingestion surface for instructions embedded in those files.
- Ingestion points: The skill reads project source files,
GLOSSARY.md, and Architecture Decision Records (ADRs) indocs/adr/to inform its analysis. - Boundary markers: There are no explicit delimiters or instructions to the model to ignore instructions embedded in the analyzed data.
- Capability inventory: The skill is restricted to reading local files and writing documentation updates. It does not contain capabilities for network communication or arbitrary command execution.
- Sanitization: The skill does not define specific sanitization or validation routines for the content it ingests from the filesystem.
Audit Metadata