gmx-v1

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.85). These links include a curl|sh to a raw GitHub install script and a direct download of a prebuilt unsigned binary from an unfamiliar GitHub user (MigOKG), plus telemetry/reporting endpoints (OKX and a Vercel app) that collect device fingerprints — together they create a high-risk vector for distributing/unwittingly running malware or leaking device identifiers and should be treated as suspicious unless you can verify the publishers and inspect the code/binary signatures.

---

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

• Third-party content exposure detected (high risk: 0.80). The skill directly fetches live price and position data from public GMX API endpoints (config.api_base_url -> https://arbitrum-api.gmxinfra.io and https://avalanche-api.gmxinfra.io) via src/api.rs get_prices/get_positions and the commands (src/commands/get_prices.rs, get_positions.rs) parse and use those values (e.g., acceptable_price/size) which can materially influence transaction construction and execution.

---

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

• Potentially malicious external URL detected (high risk: 0.80). The skill's pre-flight steps fetch and execute remote code at runtime—e.g. curl -fsSL https://raw.githubusercontent.com/okx/onchainos-skills/main/install.sh | sh and curl -fsSL https://github.com/MigOKG/plugin-store/releases/download/plugins/gmx-v1@0.1.0/gmx-v1-${TARGET} -o ~/.local/bin/gmx-v1—which are required installs and therefore constitute remote code execution from external URLs.

---

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I scanned the entire skill prompt for literal, high-entropy values that could be used as active credentials.

Findings:

• The script contains an embedded base64 string: 'OE9nNWFRUFdfSVJkektrMExOV2RNeTIzV2JibXo3ZWNTbExJUDFIWnVoZw==' which is decoded into the variable _K and then used to compute an HMAC signature (HMAC_SIG) for reporting to OKX. This is a high-entropy, literal secret key (albeit obfuscated via base64) and is therefore a real embedded credential.

Ignored items (not flagged) and why:

• All listed contract addresses and token addresses are public on-chain addresses (not secrets).
• Example commands, parameters, and placeholders (e.g., 0xYourAddress, --dry-run, documentation examples) are documentation values and not secrets.
• Simple/setup values or descriptive strings in examples are low-entropy and labeled examples, so they were not flagged.

Because the base64 value decodes to a usable key that is directly present in the code and used to sign reports, it meets the definition of a secret and should be treated as a hardcoded credential that needs remediation (remove or rotate, move to secure storage).

---

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

• Direct money access detected (high risk: 1.00). The skill is explicitly designed for crypto financial operations on GMX V1. It provides concrete, purpose-built commands to execute on-chain financial transactions: token swaps, buying/selling GLP (liquidity provision), opening/closing leveraged perpetual positions, and ERC-20 approvals. Write operations are submitted via onchainos wallet contract-call (i.e., signing and broadcasting transactions) and include execution-fee handling and contract addresses. Even though user confirmation is required, the skill clearly grants direct on-chain transaction execution capabilities.