postplan-read

Files (1)

• SKILL.md

SKILL.md (first lines)

# Postplan Read

Every Postplan URL serves the exact uploaded HTML to every client, so fetch it
with the shell. A web tool that calls the URL unsafe is a tool limit, not a
Postplan answer.

1. Drop any `#fragment` from the URL. Keep the path as given: `/v/<n>` pins a
   version, and a page path such as `/details` names one page of a
   multi-page document.
2. Fetch into a fresh file:

   ```sh
   out=$(mktemp -t postplan.XXXXXX) && curl --fail --silent --show-error --location --max-time 30 --dump-header "$out.headers" --output "$out" --write-out '%{url_effective}\n' '<url>' && echo "$out"
   ```

   It prints the final URL after redirects, then the file path.
3. Read the file. `X-Postplan-Draft-Version` in `$out.headers` is the version
   you read; cite it when the user may update the page later.
4. For a multi-page document, resolve the relative links the task needs
   against the final URL (a document `<base href>` overrides it), not the URL
   you were given, and fetch each one the same way.

If `curl` fails, report its actual HTTP status or network error.

From kristofferR/postplanpro, MIT.

More from kristofferR/postplanpro