The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts to automate web auditing tasks. It constructs command pipelines using curl, grep, sed, and awk. These commands incorporate user-provided variables (e.g., $SITE) and data extracted from target websites, which are executed in the local shell environment.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external websites, which creates a vulnerability surface.
Ingestion points: The script references/checks.md fetches content from the target origin, including robots.txt, the homepage HTML, and XML sitemaps.
Boundary markers: Absent. There are no instructions to the agent to treat the fetched content as untrusted or to ignore embedded instructions.
Capability inventory: The skill uses curl to perform status checks on URLs discovered within the fetched data and uses python3 -c to parse XML and manipulate strings.
Sanitization: The skill uses Python's urllib.parse to normalize discovery paths, but it does not fully sanitize the resulting strings against shell option injection (e.g., URLs starting with dashes) before passing them to curl in shell loops.
[DATA_EXFILTRATION]: The skill performs network operations using curl to external, non-whitelisted domains. While this is the primary purpose of the tool, it represents a network communication capability that processes external data.

sitemap-discovery-audit