customer-onboarding
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a local Python script (
scripts/check-output.py) for validating the quality of generated onboarding plans. The agent is instructed to execute this script to ensure minimum content standards are met, which is a standard functional component for this skill's purpose. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external customer and deal information—such as goals, deal notes, and contact details—to populate onboarding templates. This creates a surface for indirect prompt injection.
- Ingestion points: Data enters the agent's context through the Sales-to-CS Handoff template in
SKILL.mdand the output template intemplates/output-template.md. - Boundary markers: The skill uses Markdown tables and checklists to structure the data, providing a degree of separation from instructions.
- Capability inventory: The skill includes file-writing capabilities (creating the deliverable) and the ability to execute the
scripts/check-output.pyscript; the script itself performs local file-read operations. - Sanitization: No specific content sanitization or instruction-ignoring delimiters are defined for the processed customer data.
Audit Metadata