The Agent Skills Directory

[Command Execution] (SAFE): The skill utilizes the Bash tool to execute Python commands and scripts for managing TCP communication with the Unreal Editor on localhost. This is the intended behavior for the skill.
[Indirect Prompt Injection] (LOW): The skill ingests data from external sources within the Unreal Engine project (e.g., blueprints, asset lists) and possesses powerful capabilities like function execution and asset creation.
Ingestion points: Data is ingested through tools like scan_blueprint, list_assets, and get_node_locations via the TCP server.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are provided in the prompt templates.
Capability inventory: The skill can execute Unreal functions via call_function, create/modify assets via build_to_asset, and run local shell commands via the Bash tool.
Sanitization: Safety is mitigated by a hardcoded allowlist of permitted functions for call_function and path validation requiring /Game/ or /Engine/ prefixes with no parent directory traversal (..) allowed.

mcp-plugin