wechat-director

Pass

Audited by Gen Agent Trust Hub on May 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/visualize.py utilizes subprocess.run to execute external binaries. It invokes the bun runtime to execute a TypeScript-based backend for image generation and utilizes the obsidian CLI to update document properties and open files in the GUI for user preview. These executions are scoped to local tools and are used to implement the skill's primary automation features.
  • [EXTERNAL_DOWNLOADS]: The skill performs network operations to interact with several third-party services. It sends requests to Google Gemini and SiliconFlow for image generation, communicates with TinyPNG for file size optimization, and uploads assets to Tencent Cloud Object Storage (COS). These operations use standard API endpoints and authenticated headers using configuration sourced from a local api_keys.json file.
  • [REMOTE_CODE_EXECUTION]: The pipeline includes an optional flow that executes a vendored script (vendor/baoyu-danger-gemini-web/scripts/main.ts) using the bun runtime. While this involves executing external code, the script is a local dependency within the skill's directory structure and is used as a fallback image generation provider.
  • [CREDENTIALS_UNSAFE]: The skill accesses a local file conductor/api_keys.json to retrieve API keys. While this involves handling sensitive credentials, the skill correctly implements externalized configuration (standard secret management) rather than hardcoding secrets, and the keys are only transmitted to the intended service providers.
Audit Metadata
Risk Level
SAFE
Analyzed
May 10, 2026, 06:25 AM
Security Audit — agent-trust-hub — wechat-director