okx-agentic-wallet

Fail

Audited by Snyk on Mar 18, 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 explicitly instructs the agent to display API keys (e.g., "show both keys to the user and ask to confirm the switch") and to embed user-provided OTP/API key values in commands (e.g., onchainos wallet verify <otp> / onchainos wallet login flows), which requires emitting secret values verbatim and thus creates an exfiltration risk.

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

  • Third-party content exposure detected (high risk: 0.90). The Pre-flight Checks explicitly instruct the agent to call the public GitHub API and download installer and checksum files from raw.githubusercontent.com / github.com/releases (e.g., the install.sh and checksums in Step 1–2), which are open/public third‑party assets the agent fetches and acts on (including executing the installer), so untrusted content could materially influence behavior.

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 agent with built-in transaction execution: it supports authenticated wallet login and management, balance queries, and — critically — token transfers and smart-contract calls (commands like onchainos wallet send and onchainos wallet contract-call). The docs describe collecting params, confirming, signing in a TEE, and executing/broadcasting transactions (showing txHash, explorer links). It even supports MEV-protection flags and payable contract calls. These are specific, purpose-built financial execution capabilities for crypto on-chain transfers and contract interactions (not generic tooling), so 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
Mar 18, 2026, 12:23 PM
Issues
4