payram-testnet-testing
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill facilitates the download and execution of a setup script (
setup_payram_agents.sh) from the developer's official domain (payram.com) using bash process substitution. - [INDIRECT_PROMPT_INJECTION]: The skill exposes an attack surface by processing untrusted external data during payment verification and webhook handling.
- Ingestion points: External data enters the agent context via the
payment_lookupfunctionality (email, transaction hashes) and incoming webhook payloads. - Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish between data and commands within these external strings.
- Capability inventory: The skill utilizes tools for payment link creation, transaction status lookups, and node synchronization monitoring.
- Sanitization: There is no evidence of sanitization or strict schema validation for the external strings processed during lookup or webhook operations.
Audit Metadata