payram-agent-onboarding
Fail
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads a setup script (setup_payram_agents.sh) from the vendor's official domain (payram.com).
- [REMOTE_CODE_EXECUTION]: The skill facilitates the execution of remote code by downloading a shell script and then running it locally. This behavior is consistent with the skill's purpose of infrastructure deployment from the vendor's source.
- [COMMAND_EXECUTION]: The skill uses shell commands like chmod +x and direct execution (./setup_payram_agents.sh) to install and configure the payment environment.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources such as the PayRam API and blockchain RPC providers, which introduces a potential surface for indirect prompt injection.
- Ingestion points: API responses from payram.com and data from blockchain RPC endpoints (e.g., Alchemy, QuickNode).
- Boundary markers: No specific delimiters or instructions are provided to the agent to treat external data as untrusted content.
- Capability inventory: The skill can execute shell scripts, interact with Docker, and perform blockchain transactions.
- Sanitization: The instructions do not define sanitization or validation routines for data received from external network endpoints.
Recommendations
- HIGH: Downloads and executes remote code from: https://payram.com/setup_payram_agents.sh - DO NOT USE without thorough review
Audit Metadata