reply-in-your-voice
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze user-provided emails, creating a surface for indirect prompt injection if those emails contain malicious instructions. However, the risk is considered safe because the skill has no functional capabilities (tools, network, or file access) that could be abused. * Ingestion points: Writing samples and email messages are included in the prompt body via SKILL.md. * Boundary markers: No delimiters or safety instructions are provided to the agent to separate the training samples from the main task instructions. * Capability inventory: The skill does not define any tools, shell commands, or script execution. * Sanitization: No sanitization is applied to the user-provided data.
- [NO_CODE]: The skill is composed entirely of markdown instructions and does not include any executable scripts or software dependencies.
Audit Metadata