The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to execute a wide variety of security tools (e.g., subfinder, nuclei, ffuf, ghauri) and local scripts provided by the author (e.g., hunt.py, lead_board.py) to automate the testing process. This is the core functionality of the methodology.
[EXTERNAL_DOWNLOADS]: The workflow involves fetching data and interacting with external targets to identify vulnerabilities. These network operations are standard for security research and do not target sensitive local user data.
[PROMPT_INJECTION]: The skill processes untrusted data sourced from targets, such as JavaScript files, HTTP response headers, and page content, creating an attack surface for indirect prompt injection. 1. Ingestion points: Recon results, mapped endpoints, and vulnerability discovery responses (Phases 1-3). 2. Boundary markers: Not present. 3. Capability inventory: Extensive command execution capabilities through numerous subprocess calls to security tools. 4. Sanitization: No specific instructions for sanitizing or escaping target-provided data before agent processing are included.

bb-methodology