pipeline-status-artifact

Create (or update) a persistent Cowork live artifact that shows the full job search pipeline in a filterable table. Every time the user opens the artifact it re-reads tracker.json, recommendations.json, and artifacts-index.json from disk, so the view is always current without re-running a skill.

---

1. Verify Career Navigator setup

Application data uses the split layout defined in references/tracker-schema.md — read it before any read or write.

Read {user_dir}/CareerNavigator/tracker.json and {user_dir}/CareerNavigator/recommendations.json to confirm the files exist.

Every column this artifact renders (id, company, role, status, priority, location, job_link, follow_up_date, artifacts) is a summary row field, so the artifact loads exactly three files and never touches applications/<application_id>.json or contacts/<company-slug>.json. That is the performance win of the split — the table stays one read per file no matter how large the pipeline gets. The row also carries application, latest_stage, latest_stage_date, notes_count, stage_count, and contact_count, so a "last stage" or "contacts" column can be added with no extra reads.

If you ever extend this artifact to render per-application detail — a stage timeline, note excerpts, contact names — it must load each application's detail_file (and contacts_file, filtered to entries whose application matches the row's application label, deduped by name) explicitly, one read per application. Reading tracker.json alone will silently produce empty timelines and blank contact cells: no error, just an artifact that looks fine and shows nothing.

If both are missing, output:

Pipeline artifact not set up: run /career-navigator:launch first to initialize Career Navigator.

---

2. Check for an existing pipeline-status artifact

Load and call mcp__cowork__list_artifacts. Look for any artifact whose name contains "pipeline" or "job search" (case-insensitive).