outreach-resolve
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation guides the agent to install the paperclipai CLI and execute the @bbengamin/paperclip-mcp-server via npm and npx. These resources are provided by the skill author and represent legitimate vendor functionality.
- [COMMAND_EXECUTION]: The agent is instructed to use the paperclipai CLI to manage run records and perform health checks via curl. These commands are integral to the skill primary purpose of coordinating the outreach pipeline.
- [DATA_EXFILTRATION]: The skill performs network requests to the Paperclip API and Twenty CRM services to sync lead data and update run status. These operations are scoped to the configured environment and are necessary for record resolution.
- [PROMPT_INJECTION]: The skill processes external contact data from the Twenty CRM, creating a surface for indirect prompt injection. The risk is mitigated by a mandatory human QA gate for all ambiguous matches. Ingestion: CRM records from Twenty. Boundaries: Human-in-the-loop approval for all fuzzy matches. Capabilities: Run record updates via CLI. Sanitization: Not explicitly detailed in instructions.
Audit Metadata