The Agent Skills Directory

[SAFE]: A comprehensive analysis across all threat categories shows no signs of malicious intent or behavior. The skill provides legitimate developer tools for the Godot engine.
[COMMAND_EXECUTION]: The skill uses the gd_run interface to execute GDScript 'agent jobs' within Godot. These scripts are provided locally within the skill's scripts/agentkit/shaders/examples/ directory and are used for scene building, parameter sweeps, and performance profiling.
[DATA_EXPOSURE]: File system access is restricted to the local Godot project environment. The register_global and shader_globals functions in gd_shaders.py modify project.godot to manage shader settings, which is standard behavior for Godot plugins.
[EXTERNAL_DOWNLOADS]: The skill does not download external code or binaries. It references sibling skills and uses local dependencies provided by the same vendor.
[INDIRECT_PROMPT_INJECTION]: The skill ingests shader code (.gdshader) and project configuration files for linting and compilation. These surfaces are managed through regex-based static analysis and Godot's internal compilation API, presenting no exploitable injection vectors in this context.
[OBFUSCATION]: All code, including shader templates and Python management scripts, is provided in clear, readable text. No hidden or encoded malicious content was detected.

scenario-godot-shaders