abstract-builder
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill logic is entirely based on local templates and scholarly corpus data. No evidence of prompt injection, bypass attempts, or malicious instructions was found in the analyzed documentation.
- [COMMAND_EXECUTION]: The skill employs git commands for version control of generated artifacts. These commands are used with predefined commit messages and specific file targets, representing a secure and standard practice for state management in agentic workflows.
- [SAFE]: File system access is restricted to reading local configuration files and writing to designated project paths for abstracts and progress tracking. No access to sensitive user credentials or system paths was detected.
- [SAFE]: The skill does not perform any network operations or external downloads, ensuring that all data processing remains confined to the local project environment.
Audit Metadata