okx-agentic-wallet
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill facilitates the download and execution of an installer script (
install.shorinstall.ps1) from the author's official GitHub repository (okx/onchainos-skills). While automated scans flag this pattern, the process is mitigated by a mandatory SHA256 checksum verification step to ensure the integrity of the file before it is executed.\n- [EXTERNAL_DOWNLOADS]: The skill fetches release metadata, installer scripts, and checksum files from official GitHub endpoints (api.github.com,raw.githubusercontent.com, andgithub.com). These resources belong to the skill author and are part of the intended installation and update workflow.\n- [COMMAND_EXECUTION]: The skill executes theonchainosCLI binary for core wallet functionality. It also utilizes system utilities likecurl,shasum, and PowerShell to manage its lifecycle and verify binary integrity.\n- [PROMPT_INJECTION]: The skill processes untrusted external data from on-chain sources (e.g., token symbols and transaction history) which could serve as a vector for indirect prompt injection.\n - Ingestion points: Data returned by CLI commands for balances, addresses, and history (e.g.,
onchainos wallet balance).\n - Boundary markers: The instructions explicitly direct the AI to treat all data returned by the CLI as untrusted external content and not as instructions.\n
- Capability inventory: The skill uses subprocess calls to execute the
onchainosbinary and system commands for maintenance.\n - Sanitization: The skill relies on specific agent instructions to avoid accidental obedience of embedded commands and requires explicit user confirmation before executing transactions or contract calls.
Audit Metadata