okx-onchain-gateway

Fail

Audited by Snyk on Mar 23, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 0.90). The skill explicitly requires taking user-provided signed transactions (the --signed-tx value) and including them verbatim in broadcast commands/requests, which forces the LLM to handle sensitive secret-like values in its output and creates an exfiltration risk.

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

  • Third-party content exposure detected (high risk: 0.80). The skill's mandatory Pre-flight Checks explicitly fetch and ingest public GitHub content (curl to https://api.github.com/repos/okx/onchainos-skills/releases/latest and raw.githubusercontent.com / github.com release assets and checksums) as part of every command run, so untrusted third‑party data from the open web is read and used to decide installs/execution.

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

  • Potentially malicious external URL detected (high risk: 1.00). The skill's Pre-flight Checks explicitly fetch and execute remote installer code at runtime (e.g., curl ... "https://raw.githubusercontent.com/okx/onchainos-skills/${LATEST_TAG}/install.sh" followed by sh /tmp/onchainos-install.sh, as well as related GitHub release URLs), so external content is downloaded, executed, and relied upon for operation.

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

  • Direct money access detected (high risk: 1.00). The skill explicitly implements on-chain transaction operations: estimating gas, simulating, broadcasting signed transactions, and tracking broadcast orders across multiple blockchains (Ethereum, Solana, BSC, Arbitrum, Polygon, etc.). It includes commands like "broadcast --signed-tx", supports MEV protection parameters (e.g., enableMevProtection: true, Jito tips), and is described as the "final mile" that takes signed transactions and sends them on-chain. These are direct crypto/financial execution functions (sending transactions that move value), not generic tooling. 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
Mar 23, 2026, 01:20 PM
Issues
4