collection-schema-v3

Fail

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: CRITICALCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The instructions guide the agent to use Postman CLI tools such as postman collection lint, postman environment lint, and postman workspace lint to validate the structure of local files.
  • [DYNAMIC_EXECUTION]: The reference material includes documentation for the Model Context Protocol (MCP) and Postman scripts. It provides examples of executing commands via uvx and defining JavaScript code for request/response hooks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes structured data from local files, which presents a potential vulnerability surface if sourced from untrusted third parties.
  • Ingestion points: Files located in postman/collections/, postman/environments/, and postman/globals/.
  • Boundary markers: The skill recommends using postman lint commands and manual inspection of diffs to identify unintended changes.
  • Capability inventory: Execution of Postman CLI commands, shell execution via uvx for MCP, and JavaScript execution for collection scripts.
  • Sanitization: Instructions include quoting variable references and sanitizing filenames to ensure portable and valid YAML structure.
Recommendations
  • Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Nov 16, 2026, 01:04 PM