wechat-article-writer

Pass

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface when extracting styles from user-provided articles or processing draft content.
  • Ingestion points: User-pasted articles in the 'Style Extraction' flow and Markdown files in the drafts/ directory.
  • Boundary markers: The instructions lack explicit delimiters or warnings to the agent to ignore instructions embedded within the processed text.
  • Capability inventory: The skill can execute local shell and Node.js scripts, perform network requests to the WeChat API, and write files to the local system.
  • Sanitization: While the upload script uses HTML escaping for content payload, the agent's processing of source text lacks specific sanitization logic.
  • [COMMAND_EXECUTION]: Automates image generation and article publishing by executing local shell (scripts/export-drawio.sh) and Node.js (scripts/upload-to-wechat.js) scripts. These scripts are invoked with parameters such as titles and file paths derived from user input.
  • [EXTERNAL_DOWNLOADS]: Suggests the use of npx wechat-article-publisher in the documentation for automated publishing, which downloads and executes a package from the well-known NPM registry.
Audit Metadata
Risk Level
SAFE
Analyzed
May 6, 2026, 12:25 PM