The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the TARGET_URL variable directly within shell command blocks to execute wafw00f and curl. This pattern is susceptible to command injection if the input is not strictly validated, allowing a maliciously crafted URL (e.g., containing ; or &) to execute arbitrary commands on the host environment.
[EXTERNAL_DOWNLOADS]: The workflow executes pip install wafw00f --quiet to install the wafw00f package from the Python Package Index (PyPI) at runtime. This involves downloading and installing third-party code during the skill's execution.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by fetching and displaying response headers from user-provided URLs.
Ingestion points: Response headers and HTTP status codes are captured via curl requests to the TARGET_URL in SKILL.md.
Boundary markers: No boundary markers or explicit instructions are provided to the agent to ignore potentially malicious instructions embedded in the target's response headers.
Capability inventory: The skill possesses capabilities for shell command execution, package installation via pip, and network operations via curl and wafw00f.
Sanitization: There is no evidence of sanitization or filtering of the response data retrieved from the remote target before it is processed or displayed in the agent context.

waf-detect