track-application

Log or update an application record. Extract as much as possible from what the user said before asking for anything.

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

Application Record Schema

An application lives across three files. Never put notes, stage_history, or contacts inside a tracker.json row.

Summary row — tracker.json → applications[]:

{
  "id": "app-<company-slug>-<role-keywords>",
  "application": "<company> — <role>",
  "company": "...",
  "role": "...",
  "job_link": "...",
  "req_id": null,
  "salary_range": "...",
  "location": "...",
  "resume_version": null,
  "date_applied": "YYYY-MM-DD",
  "status": "applied | phone_screen | interview | offer | accepted | rejected | withdrew | ghosted",
  "detail_file": "applications/app-<company-slug>-<role-keywords>.json",
  "contacts_file": "contacts/<company-slug>.json",
  "notes_count": 0,
  "stage_count": 1,
  "contact_count": 0,
  "latest_stage": "applied",
  "latest_stage_date": "YYYY-MM-DD",
  "follow_up_date": "YYYY-MM-DD",
  "next_step": "...",
  "priority": "high | medium | low | null",
  "referral": { "name": "...", "relationship": "colleague | friend | recruiter | other" },
  "offer": {
    "base": "...",
    "bonus": "...",
    "equity": "...",
    "benefits_notes": "...",
    "deadline": "YYYY-MM-DD",
    "decision": null
  },
  "outcome": "pending | hired | rejected | withdrew",
  "outcome_notes": "...",
  "artifacts": []
}

Detail file — CareerNavigator/applications/<application_id>.json, one per application:

{
  "schema": "application_detail_v1",
  "application_id": "app-<company-slug>-<role-keywords>",
  "application": "<company> — <role>",
  "company": "...",
  "role": "...",
  "stage_history": [
    {
      "stage": "...",
      "date": "YYYY-MM-DD",
      "notes": "...",
      "interview_type": "recruiter | hiring_manager | technical | panel | executive | final | null",
      "interviewers": [],
      "post_notes": null
    }
  ],
  "notes": [
    { "date": "YYYY-MM-DD", "text": "..." }
  ]
}

Contacts file — CareerNavigator/contacts/<company-slug>.json, one per company, shared by every application at that company:

{
  "company": "...",
  "contacts": [
    {
      "name": "...",
      "application": "<company> — <role>",
      "title": "...",
      "email": null,
      "linkedin": null,
      "relationship": "recruiter | hiring_manager | referral | internal_contact | other",
      "notes": "...",
      "interactions": [
        { "date": "YYYY-MM-DD", "type": "call | email | linkedin | in_person", "notes": "..." }
      ]
    }
  ],
  "contact_count": 0
}

offer and referral are omitted from new records unless relevant data is present. outcome defaults to "pending" on all new records. contacts_file and contact_count are omitted until the application has a contact. previous_ids appears on a row only after a rename.

Edit every one of these files by loading and re-dumping JSON programmatically (json.load / json.dump, indent=2, ensure_ascii=False), then reload to verify. Hand-editing raw JSON has corrupted files here before.

---