Open-source x402 security testing harness (20 tests, 7 attack categories)

[msaleme]

Built the first open-source security testing harness for x402. 20 tests across 7 categories targeting the V2 attack surface:

Test Categories:

• Payment challenge validation (missing headers, malformed auth, currency mismatch)
• Recipient address manipulation (dynamic payTo routing attacks, address spoofing)
• Session token security (token fabrication, expiry bypass, sensitive data leakage)
• Spending limit exploitation (rate limit bypass, underpayment, budget exhaustion)
• Facilitator trust (fake facilitator injection, verification bypass)
• Information disclosure (leaked keys, stack traces in error responses)
• Cross-chain confusion (wrong network, wrong token type)

Zero external dependencies. Does not spend real money. Python 3.10+.

pip install agent-security-harness
agent-security test x402 --url https://your-x402-endpoint.com
agent-security list x402

Repo: https://github.com/msaleme/red-team-blue-team-agent-fabric
PyPI: https://pypi.org/project/agent-security-harness/

The harness also includes an Agent Autonomy Risk Score (0-100) that answers: how dangerous is it to let an autonomous agent pay this endpoint without human oversight?

Would welcome feedback from the x402 team on coverage gaps or additional attack vectors worth testing.

[msaleme]

Update: We've published a full AIUC-1 crosswalk mapping the x402 security tests (and all 209 tests) to the new AI agent certification standard. AIUC-1 requires quarterly third-party adversarial testing - the x402 harness maps directly to requirements D003 (restrict unsafe tool calls) and D004 (third-party testing of tool calls).

For anyone implementing x402 payment flows and evaluating AIUC-1 readiness, the framework covers payment authorization bypass, receipt manipulation, overpayment exploitation, and agent autonomy risk scoring.