self-media-wechat-publisher

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of the '@wenyan-md/cli' npm package and suggests adding a supplementary skill from a third-party GitHub repository (caol64/wenyan-cli). It also references documentation from yuzhi.tech. These sources are not in the trusted vendor list.
  • [CREDENTIALS_UNSAFE]: The skill utilizes high-privilege WeChat API credentials (AppID and AppSecret). These are managed through environment variables or specific CLI configuration commands, minimizing direct exposure within the agent context.
  • [COMMAND_EXECUTION]: Shell commands are executed to run the 'wenyan' CLI tool for content rendering and API interaction. These commands incorporate user-defined file paths and metadata fields which are not explicitly sanitized.
  • [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection by processing external Markdown documents.
  • Ingestion points: Markdown frontmatter fields (title, author, cover) and article body content in user-provided files (as shown in assets/wechat-article-template.md).
  • Boundary markers: There are no defined delimiters to isolate user-supplied strings from shell command arguments in SKILL.md instructions.
  • Capability inventory: The skill can read local files and images, execute system commands via the shell, and perform network operations through the publisher tool.
  • Sanitization: No explicit sanitization or validation logic is described for handling potentially malicious inputs within the article metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 11:04 AM
Security Audit — agent-trust-hub — self-media-wechat-publisher