opencode-api

Fail

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCREDENTIALS_UNSAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill is explicitly designed for full shell access and arbitrary command execution across Windows, macOS, and Linux, including requests for sudo privileges and system-level configuration changes.
  • [REMOTE_CODE_EXECUTION]: Instructions and templates facilitate the execution of code on remote systems via a REST API, specifically encouraging the use of Base64 encoding to obfuscate and execute complex shell commands through the pattern echo $ENCODED_CMD | base64 -d | sh.
  • [EXTERNAL_DOWNLOADS]: Contains hardcoded external IP addresses (44.197.31.152 and 52.3.242.251) designated as known endpoints for remote command execution and data transfer.
  • [DATA_EXFILTRATION]: Provides capabilities to read sensitive file contents and search for patterns across the file system, combined with tools to transmit this data to remote API endpoints (e.g., via the /session/:id/message endpoint).
  • [CREDENTIALS_UNSAFE]: Includes a 'Database Query via AI' template containing hardcoded example credentials ('user admin password admin').
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 2, 2026, 07:10 PM
Security Audit — agent-trust-hub — opencode-api