omega-cursor-cli

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by bridging user input to an autonomous agent.
  • Ingestion points: The prompt text is accepted as a positional argument in scripts/ask-cursor.mjs and passed to the downstream agent.
  • Boundary markers: The wrapper does not utilize delimiters or specific instructions to isolate the prompt from the agent's internal control logic.
  • Capability inventory: The target cursor-agent has high privileges within the workspace, including the ability to read and write code and execute shell commands.
  • Sanitization: Input data is not sanitized or validated before being passed to the execution environment.
  • [EXTERNAL_DOWNLOADS]: The skill uses npx -y to download the @cursor/agent package from the npm registry in scripts/ask-cursor.mjs and scripts/verify-setup.mjs. This facilitates access to the well-known Cursor service tools.
  • [COMMAND_EXECUTION]: The wrapper script utilizes child_process.spawn to manage local process execution. It includes platform-specific logic for Windows to resolve executable paths and uses taskkill for process lifecycle management. Shell execution is disabled to prevent common injection vectors.
  • [REMOTE_CODE_EXECUTION]: The integration with npx -y @cursor/agent allows for the dynamic loading and execution of remote code from the npm registry.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 04:33 AM