pricewin-hotel-deal-finder

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructs the agent to override its default browser capabilities (such as browser_navigate or browser_click) in favor of the provided search.js script to bypass anti-bot protections.\n- [EXTERNAL_DOWNLOADS]: The installation process (install.sh) downloads the patchright package from the NPM registry and a Chromium browser binary from the driver's official host.\n- [COMMAND_EXECUTION]: The skill executes local Node.js scripts to manage a browser daemon and perform automated searches, utilizing the Patchright library to drive Chromium.\n- [DATA_EXFILTRATION]: The skill performs network egress to booking.com, agoda.com, google.com, api.opentravel.one, and open.er-api.com to collect search results and currency exchange data.\n- [PROMPT_INJECTION]: The skill manages a surface for indirect prompt injection via scraped hotel data.\n
  • Ingestion points: Third-party hotel listing text from Agoda, Booking.com, and Google Hotels.\n
  • Boundary markers: Instructions in SKILL.md identify scraped content as untrusted third-party reference data.\n
  • Capability inventory: The skill has the ability to execute local scripts, perform network requests, and write to a local cache directory.\n
  • Sanitization: The sanitizeText() function in bin/search.js removes control characters, zero-width characters, and markdown formatting from scraped text before it is presented to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 08:49 AM
Security Audit — agent-trust-hub — pricewin-hotel-deal-finder