wechat-article-writer
Warn
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: MEDIUMINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONSAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes a feature for extracting writing styles from user-provided sample articles. This introduces a vulnerability surface where malicious instructions within the ingested text could be processed by the agent.\n
- Ingestion points: Sample articles provided by the user for style analysis (SKILL.md Section IV).\n
- Boundary markers: The instructions do not define specific delimiters or security guardrails for the agent to distinguish between data and instructions in the ingested text.\n
- Capability inventory: The skill has the ability to write files, execute scripts (bash and Node.js), and perform network operations via the WeChat API.\n
- Sanitization: There is no explicit sanitization or filtering of the sample content before analysis.\n- [EXTERNAL_DOWNLOADS]: The skill suggests the use of the external package 'wechat-article-publisher' via 'npx' without version pinning. This introduces a supply chain risk where unverified code could be downloaded and executed from the NPM registry at runtime.\n
- Evidence: SKILL.md recommends 'npx wechat-article-publisher' in the automatic upload workflow.\n- [COMMAND_EXECUTION]: The skill automates workflows by executing local scripts and CLI tools, which is necessary for its functionality but grants significant execution capabilities.\n
- Evidence: The agent is directed to run 'bash scripts/export-drawio.sh' for image conversion and 'node scripts/upload-to-wechat.js' for API interactions.\n- [SAFE]: The skill handles sensitive credentials (AppID and AppSecret) using .env files, which follows standard industry safety practices for local development environments.
Audit Metadata