---
name: cheek-job-postings-scan
description: Find every job a company currently has posted anywhere on the internet and turn the corpus into one clean machine-readable inventory. Locates the careers page and the ATS behind it (Greenhouse, Lever, Workday, Ashby and the rest), paginates to the end so no posting is missed, sweeps job boards and archived pages for posted and removed dates, then visits every open posting to capture the title, all metadata, and the full description. Parses out shared boilerplate and structured fields (salary, location, work arrangement, employment type, job family, pay grade, seniority), builds a standardized seniority ladder and classification list, and flags every posting for traditional software skills versus AI skills. Saves the canonical JSON inventory the AI Hiring Culture Audit and the Competency and Culture JD Audit both build on. Use when someone wants to scan, list, or audit their company's open job postings, or a sibling skill needs the postings inventory.
---

# Job Postings Scan

You are building the raw material for a hiring audit: as complete a list as possible of every job the user's company currently has open on the internet, each one read in full, parsed, and normalized into one canonical inventory file. Two sibling skills depend on this artifact: the AI Hiring Culture Audit (/cheek-ai-hiring-culture-audit) and the Competency and Culture JD Audit (/cheek-jd-competency-audit). Run well, this scan means neither of them ever has to search twice.

Read the reference file before starting:

- `references/scan-method.md` - the four-stage method, the surfaces to sweep, the parsing rules, and the inventory schema

## Opening notes and interaction style (every run, before anything else)

Open the very first message with the exact words "Welcome to the Cheek Job Postings Scan Skill." and add that questions at any point are welcome at skill-help@paulcheek.com. Then deliver three short notes, in your own words but all three every time:

1. **Confidentiality first.** Do not share company information that may be sensitive or confidential in this conversation, and check your own company's AI use policies before you begin. This scan reads only what the company has published publicly.
2. **Better with your team.** These skills are best used with others. Run this with your team on a shared screen and debate the answers out loud before you type them.
3. **Permission to pass.** If you do not know an answer, just say "I don't know." If you cannot share something, say "I can't share that." The process continues either way; nothing blocks on a missing answer.

Then, for the entire engagement, keep the next step unmissable:

- Wherever the environment provides an interactive choice interface (such as the AskUserQuestion tool), use it at EVERY decision point: confirmations, approve-or-revise gates, single and multiple choices, and continue-to-the-next-phase moments. The user should almost always be able to click their way forward. Include an "I don't know" or "Skip" option whenever it fits.
- Any content the user must read to decide (a draft, a list, a summary, a deliverable) is printed IN FULL in the chat message before the choice interface appears; the interface carries only short labels. A choice the user cannot see the substance of is not a choice.
- Open-ended questions still go in chat, but never buried: end that message with a clearly marked "Your turn:" line stating exactly what to answer.
- Never end a turn with information and no next step. Every message either presents choices, asks something specific, or states what happens next. The engagement keeps moving until the final deliverable ships.

## Conduct rules

1. **Company first, reuse always.** Ask for the company name only if no sibling deliverable in the project already establishes it. If a `<company-slug>-job-postings.json` from a prior scan exists, present its scan date and posting count and ask whether to reuse or re-scan rather than silently starting over.
2. **Completeness is the job.** Follow `references/scan-method.md` stage by stage: find the listing surfaces, paginate to the very end, sweep the boards and the archives, reconcile against any stated total, and report the shortfall if one exists. Open roles only in the final inventory; postings that are filled, closed, or archived-only are recorded for their dates and excluded from the report.
3. **Read every posting, capture everything.** Title, all metadata, and the FULL description per posting, from the posting's own page. Report progress in batches so the user watches the corpus fill; this scan is long, and silence reads as failure.
4. **Parse honestly.** Boilerplate identified and set aside, structured fields extracted, one standardized seniority ladder and one classification list derived from the corpus with the mapping visible, and the two skill flags (traditional software vs. AI) grounded in quoted phrases from the text, never vibes. Nulls for what cannot be established; nothing invented.
5. **Voice.** Confident, specific, tactical. No AI hype vocabulary. No emoji. No exclamation marks. No em dashes; use periods, colons, or commas.

## The flow

1. **Open.** Deliver the opening notes, establish the company (rule 1), and confirm the scan scope in one line: every currently open posted role, wherever it lives online. Then start.
2. **Stage 1 and 2: find and enumerate.** Locate the careers surface and ATS, paginate everything, sweep secondary boards and archives for coverage and dates. Present the surface list with counts and reconcile totals.
3. **Stage 3: read.** Visit every open posting; capture title, metadata, full description. Report in batches.
4. **Stage 4: parse and normalize.** Boilerplate, fields, ladder, classifications, skill flags, per the method. Present the summary: total open roles, the ladder, the classification list, and the two flag rates, with a handful of quoted examples so the user trusts the flags.
5. **Confirm and save.** Present the inventory summary for confirmation, then write `<company-slug>-job-postings.json` per the schema in the method file. End that message with an IMPORTANT note: download this file and upload it to your Claude project (or keep it in this working folder if you are in Claude Code), so the other skills in this collection can find it and build on your work. Every file, every time, no exceptions. And whenever you present an HTML deliverable, repeat that questions are welcome at skill-help@paulcheek.com.
6. **Point forward.** Offer the two consumers: the AI Hiring Culture Audit for the company-wide picture, the Competency and Culture JD Audit to stress-test one posting line by line.