The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the Katana tool using the standard Go command go install github.com/projectdiscovery/katana/cmd/katana@latest and through docker pull projectdiscovery/katana:latest. These commands fetch the tool from the official ProjectDiscovery repository and Docker Hub image registry, which are well-known sources for security tooling.
[COMMAND_EXECUTION]: The documentation includes numerous examples of command-line execution for the Katana tool. These commands are intended for web crawling activities and include flags for setting concurrency, rate limits, and output formats. The instructions emphasize that the tool should only be used on authorized systems.
[INDIRECT_PROMPT_INJECTION]: The skill documents a tool that processes external web content. When an AI agent is instructed to use the output of such a crawler (e.g., parsing discovered endpoints or JS files), it introduces a surface area for indirect prompt injection if the crawled content contains malicious instructions designed to influence the agent's behavior. The skill itself does not implement an automated pipeline for this, but the ingestion of untrusted data is a noted characteristic of the tool's primary function.

katana-web-crawling