← ClaudeAtlas

new-joblisted

Start tracking a new role from a job link or pasted description. Use when the user shares a job posting, a URL, or says they want to apply somewhere. Sets up the folder, researches the company, and reports fit before any CV is written.
ElliotJLT/boulot-os · ★ 1 · AI & Automation · score 74
Install: claude install-skill ElliotJLT/boulot-os
# Start a new application ## Step 1: Get the job description If given a **URL**, call `boulot_fetch_job` first, always. Ashby, Greenhouse, Lever and Workable publish the description as structured JSON, and about half of this vault's applications come from those boards. Fetching the page instead returns a JavaScript shell and a title. It will also tell you if a posting has been closed or unlisted, which is a different problem from being unable to read it, and worth relaying accurately. Only if that fails, try `WebFetch`. If that also fails, say so plainly and ask for the text: > I could not read that page, a lot of job boards block it. Paste the > description and I will carry on. Never guess at a job's contents from its URL, and never half-fill a folder from a failed fetch. An application built on an invented JD is worse than none. If given **pasted text**, use it directly. ## Step 2: Set up the folder ### Never stop without creating the folder Some postings do not name the company. An internal referral, a pasted extract, a careers page that assumes you know whose it is: all normal, and none of them a reason to abandon the run. If you cannot find a company name, do not ask and wait. Ask and keep going: make the folder from the clearest thing you do have, in this order, and say plainly at the end which one you used and that it can be renamed. 1. A company name in the text or the URL. 2. The product or team, if the posting names one. 3. The role plus the source, like