rpg-system-architect
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill is instructed to read external project files like
IMPROVED_PLAN.mdto guide its design. This creates a surface for indirect prompt injection if those files contain untrusted or malicious content, as there are no explicit boundary markers or sanitization steps defined. - Ingestion points:
IMPROVED_PLAN.md(via Read tool). - Boundary markers: Absent (no delimiters or 'ignore instructions' warnings for external file content).
- Capability inventory: Read, Write, Edit, Glob, Grep (capabilities to modify project files).
- Sanitization: Absent.
- [COMMAND_EXECUTION] (SAFE): The skill uses standard agent tools (Read, Write, Edit, Glob, Grep) for project management and file creation within a defined directory (
game-system/). No suspicious or arbitrary system command execution was found. - [DATA_EXFILTRATION] (SAFE): No network operations or references to external domains were detected. The skill's operations are confined to the local filesystem.
Audit Metadata