skills/mukul975/anthropic-cybersecurity-skills/performing-web-application-firewall-bypass/Gen Agent Trust Hub
performing-web-application-firewall-bypass
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The workflow in SKILL.md defines multiple command-line operations using curl, wafw00f, and sqlmap for the purpose of fingerprinting and testing target web applications. These commands are intended to be executed against external targets during authorized security assessments.\n- [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface where untrusted data from external web responses could influence the agent.\n
- Ingestion points: In scripts/agent.py, the _send method retrieves status codes and response bodies from user-defined target URLs.\n
- Boundary markers: There are no specific delimiters or instructions defined to prevent the agent from processing malicious instructions if they were contained within the WAF's response or block page.\n
- Capability inventory: The skill utilizes network requests (requests library) and suggests the use of various CLI security tools.\n
- Sanitization: The responses from the target URLs are not sanitized or filtered before the metadata (status code, length) is extracted and reported.
Audit Metadata