The Agent Skills Directory

[COMMAND_EXECUTION]: The automation scripts scripts/agent.py and scripts/process.py execute the nikto binary using the subprocess module. Commands are constructed using lists rather than shell strings, which effectively prevents shell injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The skill documentation and scripts reference the nikto scanner, a well-known open-source security tool. It requires the tool to be pre-installed on the system (e.g., via apt install nikto).
[PROMPT_INJECTION]: The skill ingests data from external web servers (headers, banners, and file paths) via Nikto scan results. This data is parsed and included in structured reports, representing a potential surface for indirect injection if the output is later processed by an LLM without proper sanitization. * Ingestion points: XML parsing logic in scripts/agent.py and scripts/process.py which extracts finding descriptions and URIs. * Boundary markers: No specific delimiters or instructions are used to separate untrusted scan data from agent instructions in the output reports. * Capability inventory: The skill is capable of performing network requests to arbitrary targets via the nikto tool. * Sanitization: scripts/process.py correctly utilizes the defusedxml library to prevent XML external entity (XXE) attacks. However, the HTML report generator in scripts/process.py interpolates finding descriptions directly into HTML templates without escaping, which could allow for cross-site scripting (XSS) if scan results contain malicious payloads.
[CREDENTIALS_UNSAFE]: The implementation guide in SKILL.md includes an example command demonstrating how to scan with authentication using the -id admin:password flag. Providing credentials directly in the command line is generally discouraged as it may expose them in process monitors and shell history.

performing-web-application-scanning-with-nikto