rollinggo-hotel-booking

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs a daily version check by fetching the SKILL.md file from the vendor's official GitHub repository (raw.githubusercontent.com/RollingGo-AI/rollinggo-hotel-skill-CN).
  • [EXTERNAL_DOWNLOADS]: The scripts/install.py script downloads the rgh binary directly from the vendor's GitHub releases (api.github.com/repos/RollingGo-AI/oauth-hotel-cli/releases/latest).
  • [COMMAND_EXECUTION]: The installation process executes npm install to set up the hotel CLI tool globally if Node.js is available.
  • [COMMAND_EXECUTION]: The skill uses Node.js and Python wrapper scripts (scripts/rgh.js, scripts/rgh.py) to execute the rgh CLI tool via spawn and subprocess.run for all booking and search operations.
  • [REMOTE_CODE_EXECUTION]: The skill downloads and executes a standalone binary executable (rgh) from the vendor's GitHub repository as part of its primary installation flow for environments without Node.js.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as hotel descriptions, facility lists, and amenity tags from CLI responses.
  • Ingestion points: Hotel search results and detail outputs from the rgh command (SKILL.md).
  • Boundary markers: None identified in the prompt interpolation templates.
  • Capability inventory: Subprocess execution and network access via the rgh binary (scripts/rgh.py, scripts/rgh.js).
  • Sanitization: No explicit filtering or sanitization of external hotel metadata is implemented before display.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 07:05 AM
Security Audit — agent-trust-hub — rollinggo-hotel-booking