The Agent Skills Directory

[SAFE]: The skill provides standard C# code templates for Unity development, including data containers, event channels, and runtime sets. No executable scripts or binary payloads are included.
[SAFE]: No network operations, credential harvesting, or unauthorized file access patterns were detected. The skill reads local project documentation files to determine architecture patterns, which is standard behavior for development-oriented agents.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local configuration files (ProjectConfig.yaml) to adjust its output logic. While this represents a theoretical attack surface for indirect prompt injection if the configuration file is compromised, the risk is minimal and handled through standard configuration reading practices.
Ingestion points: docs/ProjectConfig.yaml (file read at runtime)
Boundary markers: None identified in the provided instructions.
Capability inventory: File creation (C# scripts), execution of the refresh_unity tool.
Sanitization: None identified; however, the data is used for simple logic branching (so-first vs di-first).

uw-scriptable-object-arch