media-writer
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a legitimate and highly structured workflow for social media content production. All identified operations, including file system interactions and sub-agent management, are consistent with its stated purpose of content orchestration.
- [PROMPT_INJECTION]: The instructions contain 'CRITICAL RULES' and '铁律' (Iron Rules). These are internal behavioral constraints designed to prevent the AI from bypassing workflow stages or proceeding without explicit user consent. They do not constitute an attempt to override system safety guidelines or jailbreak the model.
- [COMMAND_EXECUTION]: The skill uses tools to read from and write to local project directories such as 'workflow/', 'archive/', and 'references/'. These actions are limited to project-related data management and do not involve execution of arbitrary system commands or access to sensitive configuration files.
- [REMOTE_CODE_EXECUTION]: No evidence of remote code execution, external script fetching, or unauthorized package management was found in the skill's instructions.
Audit Metadata