helix-mcp-craft-gtm
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill's operating principles include directives that override conflicting instructions and prohibit non-English output to maintain quality standards.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as metrics, URLs, and interview transcripts. Ingestion points: Workflow steps utilize customer-supplied data. Boundary markers: The skill requires verification of all facts and explicit assumption flagging. Capability inventory: The agent uses web tools and MCP-specific functions. Sanitization: Strict source hierarchies and verification protocols are mandated.
- [COMMAND_EXECUTION]: The skill provides a CLI command for MCP server registration.
- [EXTERNAL_DOWNLOADS]: References to a vendor-owned npm package and GitHub repository are included.
Audit Metadata