godot-dialog-systems
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill primarily consists of architectural patterns and example GDScript code for the Godot Engine. These examples focus on standard game development tasks such as dialogue management, UI effects, and resource handling within a sandbox environment. All code provided is illustrative and does not perform sensitive system operations.
- [METADATA_POISONING]: The YAML frontmatter and agent-specific configuration files (claude-code.yaml, openai.yaml) correctly scope the skill's purpose and permissions without deceptive instructions.
- [INDIRECT_PROMPT_INJECTION]: While the skill describes processing dialogue text which could contain user-provided strings, the logic is confined to UI display (RichTextLabel) and local game state (flags), posing no risk to the agent's integrity.
Audit Metadata