wechat-article-to-markdown

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches HTML content and images from WeChat's official domains (mp.weixin.qq.com and weixin.qq.com) based on user-provided URLs. This is the core functionality for converting public articles.
  • [COMMAND_EXECUTION]: The skill includes a standalone Python script, scripts/wechat_article_pipeline.py, which is intended to be executed via the command line to perform the article conversion and image downloading.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted external content from WeChat articles.
  • Ingestion points: Data enters the system via requests.get calls in WeChatArticlePipeline.fetch_html and MarkdownImageDownloader.download within scripts/wechat_article_pipeline.py.
  • Boundary markers: No explicit delimiters or instructions are used to mark the boundaries of the external content or to warn the agent to ignore embedded instructions in the generated Markdown.
  • Capability inventory: The script has the capability to read from the network and write files to the local filesystem.
  • Sanitization: While the skill sanitizes HTML tags and removes 'noise' (like WeChat-specific metadata), it does not perform sanitization or filtering to detect or neutralize natural language instructions intended for an LLM.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 03:53 PM