skills/tenequm/skills/mpp/Gen Agent Trust Hub

mpp

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The Machine Payments Protocol (MPP) is designed to process untrusted data from external network sources via HTTP headers (WWW-Authenticate, Authorization) and JSON-RPC/WebSocket messages. This constitutes an indirect prompt injection surface where a malicious party could attempt to influence agent behavior through crafted headers.\n
  • Ingestion points: SKILL.md and references/protocol-spec.md describe the handling of payment challenges and credentials from network sources.\n
  • Boundary markers: The protocol implements a cryptographic HMAC-SHA256 binding for challenge IDs to prevent manipulation of payment parameters such as amount, currency, and recipient.\n
  • Capability inventory: The skill enables agents to sign blockchain transactions, make requests to payment gateways (Stripe, Lightning), and manage state stores.\n
  • Sanitization: All payment data is base64url-encoded, and the protocol strictly requires TLS 1.2+ for all communications to ensure transport security.\n- [EXTERNAL_DOWNLOADS]: The skill documentation recommends the installation of various official SDKs and libraries to implement the protocol.\n
  • Evidence: Mentions of npm install mppx, pip install pympp, and cargo add mpp.\n
  • Source: These packages are published by the protocol maintainers and trusted ecosystem partners on public registries.\n- [COMMAND_EXECUTION]: The documentation provides examples of CLI commands for the mppx tool, which are used to manage accounts and validate server implementations.\n
  • Evidence: Examples include npx mppx validate and npx mppx account create.\n
  • Context: These are legitimate developer utility commands used for protocol testing and configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 11:13 PM