fusion-discover-skills
Pass
Audited by Gen Agent Trust Hub on Mar 21, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill's primary function is information retrieval and discovery. It uses trusted communication channels such as Fusion MCP and official GitHub APIs to interact with vendor-owned repositories.
- [COMMAND_EXECUTION]: Employs standard developer utilities like the GitHub CLI (
gh) and thenpxpackage runner for cataloging and searching. These commands are documented with safe, read-only examples to prevent accidental environment modification. - [SAFE]: Comprehensive safety guidelines are embedded within the instructions, specifically cautioning against remote code execution patterns and the execution of fetched shell content.
- [SAFE]: The skill processes data from GitHub (search results and markdown files), which represents an indirect prompt injection surface. This is addressed by instructions that limit the agent's actions to reporting and guidance, prohibiting the execution of any commands found within the discovered data. Ingestion points: GitHub search results and SKILL.md content. Boundary markers: Explicit source labeling instructions. Capability inventory: mcp_fusion_skills tool, gh CLI, and npx. Sanitization: Implicitly handled by read-only constraints and execution prohibitions.
Audit Metadata