Lp3
Medium
- Category
- MCP Least Privilege
- Confidence
- 90% confidence
- Finding
- The skill advertises direct use of shell commands and file read/write capable tools, but the manifest shown does not declare any permissions or constraints. That creates a trust and containment gap: an agent may invoke local tooling to read or modify files or execute commands without an explicit permission model, increasing the chance of unintended filesystem access or command execution. In this skill's context, those capabilities are operationally relevant for image generation, batch processing, and CLI design tools, which makes the omission more dangerous rather than less because the powerful behaviors are normalized by the workflow.
