okx-dex-bridge

Fail

Audited by Snyk on Apr 29, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E004: Prompt injection detected in skill instructions.

  • Potential prompt injection detected (high risk: 0.90). The prompt instructs the agent to silently fall back to and use static chain/bridge data (and to "not show the error to the user") when the CLI fails, which is a hidden/deceptive behavior that diverges from the skill's stated live-query purpose.

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

  • Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and ingests public third-party content (e.g., the preflight step downloads release data and installer scripts from the GitHub API / raw.githubusercontent.com in _shared/preflight.md) and repeatedly relies on untrusted CLI/API outputs (onchainos cross-chain quote/bridge/probe/status responses) which the agent must read and act on to choose routes, approvals, and execution — enabling indirect prompt-injection via those external responses.

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 designed to move cryptocurrency assets. It exposes targeted cross-chain crypto operations (quotes, prepare calldata, and — crucially — an execute command that broadcasts transfers) and includes wallet/authorization flows (approve/confirm-approve/skip-approve, --wallet, MaxUint256/unlimited approvals, polling tx status, order IDs, source/destination TX hashes). It even supports forced broadcasting (--force), approval tx submission, and full lifecycle status tracking. These are specific blockchain transaction primitives (wallet signing/authorization, broadcasting transactions, token approvals, cross-chain swaps/bridges) — not generic tooling. Therefore it grants direct financial execution authority for crypto/blockchain transfers.

Issues (4)

E004
CRITICAL

Prompt injection 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
CRITICAL
Analyzed
Apr 29, 2026, 11:52 AM
Issues
4