xrk-crawl
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a structured internal API for crawling (
runWebFetch) and browser automation (PlaywrightAgentSession). No malicious commands or instructions were identified. - [DATA_EXFILTRATION]: While the skill enables network fetching, it explicitly incorporates an SSRF protection layer via
ssrf-guard.jsand theassertUrlSafeForFetchfunction, which is designed to prevent unauthorized access to internal network resources. - [PROMPT_INJECTION]: The content consists entirely of technical documentation and API specifications. There are no attempts to override model behavior, bypass safety filters, or exfiltrate system instructions.
- [REMOTE_CODE_EXECUTION]: The skill references standard internal libraries and a well-known browser automation tool (Playwright) without downloading unverified scripts or executing arbitrary remote code.
Audit Metadata