skills/luxonis/skills/luxonis-app/Gen Agent Trust Hub

luxonis-app

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Uses the oakctl utility for project management, device interaction, and running application scripts.
  • [EXTERNAL_DOWNLOADS]: Accesses official Luxonis documentation and Model Context Protocol (MCP) resources from docs.luxonis.com and mcp.luxonis.com to ensure accurate API usage.
  • [PRIVILEGE_ESCALATION]: Includes explicit guardrails that prevent the agent from using sudo, updating firmware, or performing factory resets without user approval.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes project-level documentation to guide development tasks. (1) Ingestion points: Reads instructions and requirements from files like docs/brief.md, docs/plans/current.md, and PROJECT_BRIEF.md. (2) Boundary markers: Does not implement explicit delimiters or warnings to ignore instructions found within these documents. (3) Capability inventory: Capable of executing shell commands via oakctl and writing to the local filesystem. (4) Sanitization: Content from workspace files is interpreted directly to inform project planning and implementation steps.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 01:16 AM
Security Audit — agent-trust-hub — luxonis-app