The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to download and run @axe-core/cli, capture-website-cli, and playwright from the npm registry. It also uses curl to fetch sitemaps and web content from external URLs for auditing purposes.
[COMMAND_EXECUTION]: The skill provides command templates to execute shell operations for web scanning, screenshot capture, and data parsing using grep. It includes the --no-sandbox flag for Chrome, which is standard for containerized environments but reduces browser isolation.
[PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection by ingesting untrusted data from the web. \n
Ingestion points: Fetches and processes external site data via curl and link crawling. \n
Boundary markers: Does not specify the use of delimiters or clear separation between retrieved web content and the agent's instructions. \n
Capability inventory: Has the ability to execute shell commands (npx), write files to the workspace (--save), and perform network requests. \n
Sanitization: Implements basic filtering of URLs based on extensions (e.g., skipping .zip, .pdf) but lacks content-level sanitization for the retrieved HTML or text.

web-scanning