skills/okx/onchainos-skills/okx-defi/Gen Agent Trust Hub

okx-defi

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides fallback instructions to download and execute installation scripts if the required CLI tool is not found. Evidence: The preflight guide in _shared/preflight.md directs the agent to execute install.sh via sh or install.ps1 via iex from the vendor's GitHub repository.
  • [EXTERNAL_DOWNLOADS]: The skill references and downloads setup scripts from the official OKX GitHub organization. Evidence: https://raw.githubusercontent.com/okx/onchainos-skills/main/install.sh and https://raw.githubusercontent.com/okx/onchainos-skills/main/install.ps1.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external DeFi protocols, which can include attacker-influenced metadata like product names or descriptions. Ingestion points: External protocol data is retrieved through CLI commands like defi search and defi detail as described in references/invest-cli-reference.md. Boundary markers: There are no explicit instructions to use delimiters or sanitization warnings when displaying these results to the agent. Capability inventory: The skill can generate and broadcast on-chain transactions via the deposit, redeem, and claim commands, allowing for the execution of financial operations. Sanitization: No specific sanitization or validation logic is defined for the external string fields displayed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 02:18 AM
Security Audit — agent-trust-hub — okx-defi