agentbox-openrouter
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands including
cat,jq, andmvto update configuration files. It also uses a vendor-specific CLIopenclawto manage service state. All commands are consistent with the skill's stated purpose. - [DATA_EXPOSURE]: The skill accesses and modifies
~/.openclaw/openclaw.json. While this file contains sensitive information like API keys, the access is restricted to local configuration management as requested by the user. No evidence of unauthorized data transmission was found. - [SAFE]: The skill does not contain any prompt injection attempts, obfuscated code, or unauthorized network operations. External URLs provided are for official documentation and account management at OpenRouter.
Audit Metadata