character-card-builder
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists of natural language instructions designed to guide an AI agent in a creative writing workflow. It does not perform any unauthorized system operations or network requests.
- [NO_CODE]: The skill is composed entirely of markdown instructions and lacks any executable scripts (Python, JavaScript, etc.), which inherently eliminates the risk of code-based attacks such as remote code execution or privilege escalation.
- [PROMPT_INJECTION]: The instructions include specific behavioral constraints (e.g., 'never writes or rewrites the author's prose', 'never invents the character for the writer') that reinforce safety and ensure the agent remains in its intended role.
- [DATA_EXPOSURE]: The skill mentions reading from a project 'Map' and manuscript via the platform's Model Context Protocol (MCP) connector. This is a functional feature of the author's ecosystem used to provide context for character development. The 'writes_back' capability is disclosed in the metadata and is used solely to store the resulting character cards within the user's workspace.
Audit Metadata