use-cli-scout

# use-cli-scout ## When to use Reach for scout whenever the answer depends on something on the web. It returns the page body itself, so the answer can rest on the primary source. When the input carries an `http://` or `https://` URL, fetch it before answering. When unsure, fetch. ## Commands ### When a URL is given `repo-overview` and `repo-read` accept a full URL as well as `owner/repo`. A Slack permalink is detected by `fetch` and routed to the Slack Web API. | URL shape | Command | | --------------------------------------------- | ------------------------------------- | | `github.com/<owner>/<repo>` | `scout repo-overview <url>` | | `github.com/<owner>/<repo>/blob/<ref>/<path>` | `scout repo-read <owner/repo> <path>` | | anything else (Slack permalinks included) | `scout fetch <url>` | ### Researching `-l ja` and `-l en` pin the search language, which is auto by default. Add `--js` for a page that renders through JavaScript, `--raw` for the whole page rather than the extracted article. | What you want | Command | What comes back | | ----------------- | ---------------------------------- | -------------------------------------- | | Candidate sources | `scout search "query"` | One URL per line | | A topic's bodies | `scout research "topic" -d <1-10>` | Markdown rep...