openclaw-chinese-documentation

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTIONNO_CODE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install the openclaw package from the NPM registry and to clone the documentation repository yeuxuan/openclaw-docs from GitHub.
  • [COMMAND_EXECUTION]: Provides detailed shell commands for platform deployment, including global package installation (npm install -g), service initialization (npx openclaw init), and container management via Docker Compose.
  • [REMOTE_CODE_EXECUTION]: The skill encourages the execution of code from external repositories and packages (openclaw via NPM and yeuxuan/openclaw-docs via Git) which are downloaded at runtime.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary purpose is to guide the agent through 767 external tutorials. This volume of third-party data creates a significant injection surface if the external documentation were to be compromised or contain adversarial instructions.
  • Ingestion points: External documentation repository github.com/yeuxuan/openclaw-docs and the live site openclaw-docs.dx3n.cn.
  • Boundary markers: None identified to separate instructions from the documentation content.
  • Capability inventory: The skill provides commands for local file writes (configuration), shell execution (CLI), and network testing (curl commands to AI providers).
  • Sanitization: No evidence of content sanitization or validation for the external documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 06:50 PM
Security Audit — agent-trust-hub — openclaw-chinese-documentation