The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process content from arbitrary external URLs, including websites, PDFs, and repositories. This untrusted content is converted to Markdown and presented to the agent, creating a surface for instructions embedded in the source material to influence the agent's actions.
Ingestion points: The scrape, crawl, map, extract, and search commands in SKILL.md ingest data from any user-provided or discovered URL.
Boundary markers: The instructions do not explicitly specify the use of delimiters or 'ignore' instructions when handling the ingested content.
Capability inventory: The agent can execute shell commands (npx webcrawl) and perform network operations to interact with the web.
Sanitization: While the tool converts content to Markdown, it does not provide filtering for natural language instructions embedded within that content.
[COMMAND_EXECUTION]: The skill uses npx to execute the webcrawl command-line utility. This is the intended primary mechanism for the skill's functionality and is used to interact with the vendor's provided tools.
[EXTERNAL_DOWNLOADS]: The skill references npx webcrawl, which involves downloading and executing packages from the npm registry. These resources are consistent with the vendor's own ecosystem and are required for the skill to operate.

webcrawl-crawl