openclaw-platforms
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches installation scripts from the vendor's official domain (openclaw.ai) and references the openclaw package on npm.
- [REMOTE_CODE_EXECUTION]: Suggests piping a remote script from https://openclaw.ai/install.sh into bash for automated setup on Linux and cloud environments.
- [COMMAND_EXECUTION]: Instructs the agent on using various CLI tools for platform management, including systemctl, launchctl, and tccutil.
- [PROMPT_INJECTION]: Documentation of the 'Canvas' feature reveals an attack surface for indirect prompt injection via canvas.navigate and canvas.eval when processing untrusted web content.
- Ingestion points: references/en/platforms/mac/canvas.md, references/en/platforms/ios.md.
- Boundary markers: Absent for remote content.
- Capability inventory: canvas.eval, system.run.
- Sanitization: Absent for remote script execution within the canvas.
Audit Metadata