skills/mukul975/anthropic-cybersecurity-skills/testing-for-xss-vulnerabilities-with-burpsuite/Gen Agent Trust Hub
testing-for-xss-vulnerabilities-with-burpsuite
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted HTML content from target websites to identify reflected payloads. These payloads are stored in a JSON report without sanitization, which creates a risk of indirect prompt injection for any AI agent that subsequently processes the findings.\n
- Ingestion points: The
scripts/agent.pyscript retrieves responses from user-defined target URLs to check for payload reflection.\n - Boundary markers: The generated report (
xss_burp_report.json) contains raw XSS payloads without delimiters or warnings to downstream agents.\n - Capability inventory: The skill can perform network requests and write results to the local filesystem.\n
- Sanitization: Untrusted data from remote targets is included in reports without escaping or sanitization.\n- [EXTERNAL_DOWNLOADS]: The documentation suggests installing the third-party tool
dalfoxfrom a non-trusted GitHub repository (github.com/hahwul/dalfox).\n- [DATA_EXFILTRATION]: Thescripts/agent.pyscript disables SSL certificate verification (verify=False) while supporting the transmission of authentication tokens via theAuthorizationheader. This configuration can lead to the exposure of credentials to network interceptors during testing sessions.
Audit Metadata