reply-draft-assistant
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and classify prospect replies from LinkedIn and email, which creates an attack surface for indirect prompt injection where malicious content in a prospect's message could influence the agent's behavior. 1. Ingestion points: Prospect messages are retrieved via the La Growth Machine MCP tools or pasted directly by the user as described in references/fetch-conversations.md. 2. Boundary markers: The instructions do not specify the use of delimiters or boundary markers to isolate untrusted external content from the skill's instructions. 3. Capability inventory: The skill has permissions to execute the send_linkedin_message and send_email_message tools after receiving user approval. 4. Sanitization: There is no mention of sanitizing or filtering input text for malicious instruction patterns.
- [EXTERNAL_DOWNLOADS]: The skill documentation and README reference external links for installation and support, including lagrowthmachine.com, app.lagrowthmachine.com, tally.so, and the vercel-labs/skills GitHub repository. These are identified as legitimate vendor resources or well-known services and do not pose a security risk.
Audit Metadata