cold-email
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is composed strictly of markdown documentation and evaluation data. It contains no executable scripts, shell commands, or external dependencies that could introduce security risks.
- [PROMPT_INJECTION]: The instructions incorporate safety guardrails that explicitly limit the agent's authority, stating that drafting emails does not permit sending or scheduling, and requires verification of privacy requirements and opt-out lists before use.
- [DATA_EXPOSURE]: The skill encourages reading a local context file (
.agents/product-marketing-context.md) to improve output relevance. This is a localized context-gathering action and does not involve the exposure of credentials, environment variables, or sensitive system configurations. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external signals (like LinkedIn activity or company news) provided by the user. However, since the skill does not utilize tools for file-writing, network communication, or code execution, the potential for an indirect injection to cause system harm is mitigated within this scope.
Audit Metadata