wechat-product-manager-writer
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
Bashtool to execute a local Python script (scripts/generate_image.py) for generating article cover images. This is a functional requirement and uses standard Python patterns to call image generation APIs. - [EXTERNAL_DOWNLOADS]: The
generate_image.pyscript makes network calls to OpenAI's and Google's official API endpoints for DALL-E and Gemini. These are well-known technology services and the communication is required for the image generation feature. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data via
WebFetchto perform research. It implements boundary markers through specific article templates and provides sanitization instructions via a detailed fact-checking guide (references/fact-checking.md) and advice on distinguishing facts from opinions. The skill has capabilities includingBashand file writing, but its instructions mitigate the risk of accidental obedience to malicious instructions in web content. - [CREDENTIALS_UNSAFE]: The skill correctly identifies that API keys should be stored in environment variables rather than being hardcoded in scripts or instructions, adhering to standard security best practices for secret management.
Audit Metadata