okx-agentic-wallet

Fail

Audited by Snyk on May 20, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 1.00). The prompt instructs the agent to accept and use sensitive values (verification codes and to display/confirm API keys) and to embed them verbatim in CLI commands (e.g., onchainos wallet verify <code> and "show both keys to the user"), which requires the LLM to handle and output secret values directly, creating exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.90). The skill's mandatory preflight (/_shared/preflight.md Step 1–3) instructs the agent to fetch live release metadata and installer scripts from public GitHub endpoints (https://api.github.com and raw.githubusercontent.com) and to download/verify and potentially execute installers based on that content, so it consumes untrusted, public third‑party data that can materially influence subsequent tool actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).


MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

  • Direct money access detected (high risk: 1.00). The skill is explicitly a crypto wallet/transaction manager for OKX Agentic Wallet. It defines concrete commands to send native and ERC‑20/SPL tokens (onchainos wallet send), perform contract calls (onchainos wallet contract-call), manage gas payment via Gas Station (setup/enable/disable/update-default-token), sign messages/transactions (TEE signing, personalSign, EIP‑712), view balances/history, and export/import actions. These are specific on‑chain financial execution primitives (wallet transfers, contract interactions, signing/broadcasting txs), not generic tools. Therefore it grants direct financial execution authority.

Issues (4)

W007
HIGH

Insecure credential handling detected in skill instructions.

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

W009
MEDIUM

Direct money access capability detected (payment gateways, crypto, banking).

Audit Metadata
Risk Level
HIGH
Analyzed
May 20, 2026, 01:22 AM
Issues
4