The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently utilizes the execute_tool_script capability of the Unreal Engine MCP server to run dynamically generated Python code.
Evidence: This pattern is implemented in actors.py, camera.py, hierarchy.py, and panels.py to handle batch actor spawning, camera movements, and UI rendering.
Context: The code generation is used to mitigate performance issues and host instability by reducing the number of JSON-RPC round-trips to the engine.
Sanitization: The skill uses double-JSON encoding (e.g., in actors.py and panels.py) to safely embed data into the generated script strings, which is a defensive measure against script injection.
[INDIRECT_PROMPT_INJECTION]: The skill processes network topology, metrics, and incident data from external sources which are then used in the 3D visualization and script generation.
Ingestion points: renderer.py and telemetry.py receive data from sources like pyATS, SuzieQ, CML, and SNMP traps.
Boundary markers: None explicitly present to separate untrusted data from the agent's internal instructions.
Capability inventory: The skill possesses extensive control over the Unreal Engine environment, including actor spawning, property manipulation, and script execution.
Sanitization: Basic character replacement is performed on hostnames in actors.py to ensure they are safe for use as actor names.
[EXTERNAL_DOWNLOADS]: The documentation references official software downloads.
Evidence: SKILL.md directs users to unrealengine.com for the Unreal Engine software.
Note: This is a neutral reference to a well-known service for necessary prerequisites.

ue5-network-viz