dsh-dead-links

No matter how carefully you write your docs, external links eventually rot—an API shuts down, a domain expires, a page moves—and every dead link quietly erodes reader trust. dsh-dead-links turns this into a single automated scan: point it at a documentation directory and it recursively walks every Markdown file, extracts http(s) links line by line, probes each URL concurrently, and returns a structured JSON report with file path, line number, and status code, so broken links are immediately visible.

Under the hood, it prefers HEAD requests to minimize bandwidth, automatically falling back to GET when a 405 or 403 is returned; concurrency defaults to 5 (configurable 1-10), and duplicate URLs are fetched once then mapped back to every occurrence. The entire process is strictly read-only: no file is modified, written, or cached to disk. Network exceptions are never thrown—each one is folded into the JSON error field, making the output safe for upstream tooling to consume.

Ideal for individuals and teams maintaining technical documentation, API references, internal wikis, or blog sites. Paired with a future scheduler for periodic triggers, it becomes a weekly dead-link audit loop: scan, push the report to a model for review or auto-fix, and keep documentation health checks as a fully read-only, structured pipeline.