The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill is primarily composed of executable shell commands and a custom bash script (recon.sh) for automating security testing. It grants the agent direct access to execute numerous networking tools including amass, nuclei, ffuf, and nmap-like scanners.
[PROMPT_INJECTION] (HIGH): The skill exhibits a major Category 8 (Indirect Prompt Injection) attack surface. It fetches data from untrusted external sources (e.g., historical URLs via waybackurls and gau) and feeds them into subsequent analysis tools.
Ingestion points: waybackurls, gau, subfinder, amass, and paramspider.py fetch data from the public internet and historical web archives.
Boundary markers: None. There are no delimiters or instructions provided to the agent to treat the output of these tools as data rather than potential instructions.
Capability inventory: The skill includes high-privilege capabilities such as automated vulnerability scanning (nuclei), directory brute-forcing (ffuf), and XSS testing (dalfox).
Sanitization: None detected. Data is piped directly from collection tools into execution/probing tools without filtering or escaping.
[EXTERNAL_DOWNLOADS] (MEDIUM): The workflow relies on over a dozen third-party security tools that are not part of a standard operating system distribution. This introduces a large supply chain risk as the user is expected to have these tools installed from various external sources (GitHub, Go registries, etc.).

Red Team Tools and Methodology