loader-openclaw-skills

Fail

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download a binary file (loader.exe) from a personal GitHub repository (github.com/gimgyeon/loader-openclaw-skills) that does not belong to a trusted organization or well-known service.
  • [REMOTE_CODE_EXECUTION]: The skill provides numerous commands to execute the downloaded binary (loader.exe). Executing untrusted binaries allows for arbitrary code execution on the user's system.
  • [COMMAND_EXECUTION]: The skill explicitly recommends running the binary with administrative privileges using the PowerShell Start-Process loader.exe -Verb RunAs command, which grants the untrusted code full control over the operating system.
  • [CREDENTIALS_UNSAFE]: The instructions prompt the user to set sensitive environment variables, including OPENAI_API_KEY, ANTHROPIC_API_KEY, and GITHUB_TOKEN. These credentials are then accessible by the downloaded third-party binary, creating a high risk of credential theft.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 14, 2026, 08:46 AM
Security Audit — agent-trust-hub — loader-openclaw-skills