x402

Warn

Audited by Socket on Aug 16, 2026

4 alerts found:

SecurityAnomalyx3
SecurityMEDIUM
SKILL.md

SUSPICIOUS: the core payment and monetization behavior matches the stated purpose, so this is not obviously deceptive, but it is a high-impact financial skill. The main concerns are autonomous real-money actions, routing through a third-party facilitator domain, and reliance on another skill plus unseen setup scripts. This looks coherent but materially risky rather than benign.

Confidence: 84%Severity: 78%
AnomalyLOW
client.py

No clear evidence of intentional malware or sabotage is present in this fragment (no dynamic execution, no reverse shells, no obvious stealth exfiltration). However, there are meaningful security risks: (1) it creates and stores an EOA private key on disk (/data/workspace/.x402/buyer.key) when missing; (2) it can install a custom trusted CA and force HTTP(S) proxying based on environment variables, which can enable MITM/interception if the environment is attacker-influenced; and (3) it writes a local payment ledger file containing metadata. Overall, treat this as a sensitive security component requiring strict environment control and secrets management.

Confidence: 66%Severity: 55%
AnomalyLOW
bazaar.py

No direct evidence of classic malware (backdoor, shell, exfiltration, exec) is present in this fragment. However, there are meaningful security risks: (1) potential SSRF/probing of attacker-influenced URLs via probe_402 without a strict allowlist, and (2) dynamic sys.path insertion/import of local modules (exports/client) which could be exploited if those directories are not secured. Overall this looks like a payment/discovery component with safety gating, but the module-loading pattern and network probing surface warrant review.

Confidence: 62%Severity: 55%
AnomalyLOW
templates/monthly.json

This fragment is a JSON configuration with no direct code execution or observable malicious logic by itself. However, it includes an admin token placeholder (`facilitator_admin_token`) and payment/settlement parameters (`pay_to`, `facilitator`, `price_usd`, and routing rules), which could enable sensitive network authentication and fund routing in the consuming application. Without the consuming code, malware cannot be confirmed, but the configuration is high-impact and warrants review of where/how these values are used (especially token handling and payment destination integrity).

Confidence: 55%Severity: 50%
Audit Metadata
Analyzed At
Aug 16, 2026, 05:02 AM
Package URL
pkg:socket/skills-sh/starchild-ai-agent%2Fofficial-skills%2Fx402%2F@479f221b4b43d911ccf273fc04c6a3292dfb6c8575e47cf4a98e26ee0befaa10
Security Audit — socket — x402