direct-mail
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions and strategic frameworks for physical mail prospecting. No code, scripts, or automated network operations are included in the skill body or referenced files.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest data from .agents/sales-context.md in the project root to personalize campaign planning. 1. Ingestion point: .agents/sales-context.md (referenced in SKILL.md). 2. Boundary markers: None provided. 3. Capability inventory: The skill itself contains no executable scripts or subprocess calls; it relies on the agent's external tools. 4. Sanitization: None provided. While this presents an attack surface for instructions embedded in local project data, it is a standard mechanism for context-aware agents and represents low risk in this configuration.
Audit Metadata