openclaw-persona-forge
Pass
Audited by Gen Agent Trust Hub on Apr 6, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (
gacha.py) to perform random character generation. This script uses thesecretsmodule for secure randomization and includes error handling for user-supplied arguments, mitigating potential command injection at the script level. - [DATA_EXPOSURE]: The skill writes temporary files to
/tmpand generatesSOUL.mdandIDENTITY.mdfiles in user-specified directories. This is legitimate behavior for its purpose of creating and exporting persona configurations. - [PROMPT_INJECTION]: The skill facilitates indirect prompt injection by design, as it interpolates user-provided preferences into templates for character profiles and external image generation prompts.
- Ingestion points: User-provided character preferences and the 'draw count' argument in
SKILL.md. - Boundary markers: Absent when interpolating character details into the final markdown files.
- Capability inventory: File system writes (Write tool), subprocess execution (Python), and potential network operations (via optional external image generation skills).
- Sanitization: The skill implements filename sanitization (restricting to alphanumeric characters and hyphens) for temporary files, reducing the risk of path traversal or shell injection.
Audit Metadata