headless-support

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the official MCP server for Plain (https://mcp.plain.com/mcp), a well-known customer support platform, to facilitate AI agent integration.
  • [COMMAND_EXECUTION]: Includes a Python script (scripts/check-output.py) used to validate the quality and presence of specific terms in the generated support plan deliverables.
  • [INDIRECT_PROMPT_INJECTION]: The system is designed to ingest and process untrusted data in the form of customer support tickets.
  • Ingestion points: Processes customer threads and inquiries (referenced in SKILL.md and references/byoai-headless-stack.md).
  • Boundary markers: Includes persona instructions such as "Answer from help center articles only" to limit model behavior, though it lacks formal structural delimiters for untrusted input.
  • Capability inventory: The agent is empowered to draft replies (addGeneratedReply), assign labels, and set ticket priorities.
  • Sanitization: No explicit sanitization or validation logic for the incoming ticket content is defined.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 06:15 PM
Security Audit — agent-trust-hub — headless-support