daily-schedule

Scheduling (Claude Cowork)

This skill is designed to run as a daily scheduled task using Cowork’s /schedule (or equivalent). Example payload the user might schedule:

Run the Career Navigator daily-schedule skill for my job search folder. Deliver the daily brief per skills/daily-schedule/SKILL.md.

After the first successful run, Cowork typically refines the scheduled prompt with learned paths, connectors, and context.

Workflow

1. Resolve {user_dir} and validate data files

Check:

• {user_dir}/CareerNavigator/profile.md
• {user_dir}/CareerNavigator/ExperienceLibrary.json
• {user_dir}/CareerNavigator/tracker.json
• {user_dir}/CareerNavigator/artifacts-index.json
• {user_dir}/CareerNavigator/StoryCorpus.json (create-on-demand is allowed via mine-stories)

If missing, output:

Daily brief skipped: run /career-navigator:launch to initialize Career Navigator.

2. Reconcile artifacts before summary

Before computing the brief, scan {user_dir} for any PDF/DOCX artifacts.

If one or more artifact files are present, run artifact-saved first so artifacts-index.json is synchronized before counts are calculated.

If no artifacts are found, continue without running artifact-saved.

2.5 Auto-ingest new source documents

Before finalizing the brief, detect whether newly discovered resume/CV-style files are present in {user_dir} but not yet represented as source entries in ExperienceLibrary.

When such files are found:

• Automatically run add-source for each newly discovered source document (no manual prompt required).
• After each successful ingest, re-run artifact-saved once to keep artifacts-index.json consistent.
• In the final brief, include a one-line note listing how many files were auto-ingested.

If an ingest fails for any file:

• Continue with the daily brief.
• List failed filenames under a short "Needs attention" line.
• Recommend rerunning /career-navigator:add-source for those specific files.

2.6 Incremental story corpus refresh

Before finalizing the brief, detect whether story-bearing source files changed since the last story mining run:

• journals / logs
• PKM notes/exports
• interview debrief notes
• resume/CV/cover-letter prose that may add interview evidence

Use CareerNavigator/StoryCorpus.json source_index entries (path + mtime) as the baseline when available.

When at least one new/changed eligible source file is detected:

• Automatically run mine-stories in incremental mode (no manual prompt required).
• Refresh StoryCorpus.json and its meta.updated + source_index.
• In the final brief, add one short line:
  • Story corpus refreshed: {n_files} source file(s) scanned, {n_stories} stories added/updated.

If refresh fails:

• Continue with the daily brief.
• Add a concise "Needs attention" line:
  • Story corpus refresh failed; run "mine-stories" to retry.

2.7 Monthly career-plan checkpoint

If {user_dir}/CareerNavigator/career-trajectory.md exists, determine whether it is stale:

• Prefer to parse the report as_of date from the career_trajectory_v1 JSON block, or a heading like ## Career trajectory analysis ({YYYY-MM-DD}).
• Consider the file "stale" when:
  • the file is missing (first run), OR
  • it was generated more than 30 days ago, OR
  • the tracker has 5 or more new outcome events since the file's as_of (rejections, offers, accepted completions; do not count pending).

If stale, set career_plan_refresh_due = true.

2.8 Offer evaluation nudge

Check tracker for any active application where:

• status is "offer"
• and {user_dir}/CareerNavigator/offer-context-{application_id}.json does not exist (offer evaluation has not been run yet)

If any exist, set offer_evaluation_due = true and capture up to 3 items with company, role, and deadline (if present).

3. Build the daily operating brief

Read {user_dir}/CareerNavigator/tracker.json and {user_dir}/CareerNavigator/artifacts-index.json.

Compute:

• Pipeline counts by status (applied, phone_screen, interview, offer, accepted, rejected, withdrew, ghosted) from tracker.json; plus count of considering records from recommendations.json (reported separately as "X roles under consideration")
• Overdue follow-up count (using follow_up_date if present)
• Meetings today — count stage_history entries where date is today (local) and stage matches the allowlist below (case-insensitive substring on stage string)
• Artifact counts by type (resume, cover_letter)

3.1 Stage allowlist — interview / recruiter / screen today

Treat a stage_history row as “meeting today” when date equals today and any substring match holds:

interview, recruiter, phone screen, phone_screen, hiring manager, hm , hm interview, technical, panel, onsite, executive, final round, final interview

Build the list of applications that have at least one such row today (dedupe by application id).

3.2 Pre-interview brief subsection

• If no applications qualify for meetings today: omit the entire Pre-interview brief subsection from the output (no placeholder).
• If one or more qualify: append a subsection after the pipeline block (see §5) titled Pre-interview brief (today). For each qualifying application, produce short bullets only (strict brevity — this is not a full prep doc):
  • One news or company hook (only if grounded via tracker/artifacts or verifiable web; else say “no verified headline—run /career-navigator:prep-interview for depth”)
  • 2–3 talking points from ExperienceLibrary + role fit
  • Interviewer reminder if contacts[] or stage_history[].interviewers has a name
  • One watch line (risk, gap, or follow-up to confirm)

Follow agents/interview-coach/AGENT.md in morning_section mode for tone and limits. You are implementing that mode inline in this skill (no separate skill file).

3.3 /career-navigator:morning-brief — focused run

If the user’s message is primarily a day-of interview request (triggers include /career-navigator:morning-brief, “morning brief before my interview”, etc.):

1. Still perform §1 validation and, when appropriate, §2 artifact reconciliation (lightweight).
2. Skip the full pipeline table and artifact counts unless the user also asked for a full digest.
3. Output only:
   • A one-line header: Career Navigator — Pre-interview brief [{today}]
   • The Pre-interview brief (today) content per §3.2, or a clear line that no qualifying meetings appear in tracker.json for today—then suggest /career-navigator:track-application to log scheduled screens and /career-navigator:prep-interview for deep prep.

If the user runs a general daily trigger, produce the full template in §5, including Pre-interview brief when §3.2 applies.

4. Add market and strategy prompts

If user has run market workflows recently, summarize top signal in one line. If no recent strategy signals exist in tracker data, append:

Run /career-navigator:suggest-roles to refresh strategy signals for job-scout ranking.

If no market brief has been generated recently, append:

Run /career-navigator:market-brief for weekly demand and displacement updates.

5. Output format

**Career Navigator — Daily Brief** [{today}]

Pipeline
  Considering      {n}
  Applied          {n}
  Phone screen     {n}
  Interview        {n}
  Offer            {n}

Follow-up due      {n}
Meetings today     {n from §3.1 or "None scheduled"}

Artifacts
  Resumes          {n}
  Cover letters    {n}

If §3.2 applies, append:

Pre-interview brief (today)
  {per-application concise bullets — see §3.2}

Then append at most 3 concise action bullets, highest urgency first. Include:

• If offer_evaluation_due: an "Offer evaluation due" bullet with the top item(s) and prompt to run /career-navigator:evaluate-offer.
• If career_plan_refresh_due: a "Career plan refresh due" bullet prompting the user to run /career-navigator:career-plan.

Guardrails

• Keep this as the primary place for routine summaries; do not duplicate in focus-career
• If data is sparse, say so explicitly and keep recommendations lightweight
• Do not fabricate market news, dates, or deadlines