editor-selection-get
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the
unity-mcp-clibinary to interact with the Unity Editor API. This is the primary intended function of the skill to fetch editor state. - [EXTERNAL_DOWNLOADS]: The documentation suggests installing the
unity-mcp-clipackage usingnpm install -g. This download targets the official npm registry, which is a well-known and trusted service for package management. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Unity Editor, such as GameObject names, paths, and asset metadata. This presents a potential surface for indirect prompt injection if an agent processes a project containing maliciously crafted object names.
- Ingestion points: Unity Editor environment (GameObject names, paths, and Asset GUIDs) via
editor-selection-gettool output. - Boundary markers: None. The data is returned as structured JSON but without specific instructions to ignore embedded text.
- Capability inventory: The skill only performs read operations and does not include capabilities for file system modification, network exfiltration, or dynamic code execution.
- Sanitization: No sanitization or filtering of object names or paths is performed before the data is added to the agent's context.
Audit Metadata