saas-health-check
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of business documents, analytics exports, and financial records which could contain hidden malicious instructions aimed at overriding agent behavior.
- Ingestion points: File reading during 'UPDATE' and 'FIND' modes from directories like 'metrics/', 'product/', 'strategy/', and 'analytics/', as well as files provided via the '@file' syntax.
- Boundary markers: Absent; there are no instructions to the agent to treat ingested data as non-executable text or to ignore embedded instructions.
- Capability inventory: The skill can search and read files, write data to the filesystem, and invoke other agent capabilities through slash commands like '/context-save' and '/present'.
- Sanitization: Absent; the instructions do not specify any validation or filtering of external content before it is processed or integrated into output.
Audit Metadata