Daily Driver

Personal productivity assistant. A durable data layer for daily planning, tracking, and job search with a Claude-powered conversational layer on top.

What it does

Daily Driver keeps the unglamorous parts of a focused workweek out of your head and in structured files:

• Workspace — one directory holds plans, notes, tracker, config. daily-driver init . scaffolds it.
• Tracker — YAML-backed list of anything you follow up on (applications, tasks, chores, contacts). Categories are config-driven and extensible via --extra key=value.
• Focus mode — file-locked toggle that suppresses scheduled check-ins.
• Status dashboard — totals, stalled items, last week of activity as a Rich table (or JSON).
• Job-board scraper — multi-source scrape (RemoteOK, WeWorkRemotely, HN Who's Hiring, HN YC-funded jobs, Greenhouse, Ashby, Lever, Workable, Workday, LinkedIn/Indeed via JobSpy, Apple), with board auto-discovery for the ATS sources, dedup, filter, and AI-driven enrichment via claude or a local Ollama model.
• Gather modules — typed readers for git history and macOS Calendar (icalBuddy).
• Integrations — thin subprocess bridges to claude, pbcopy, launchctl.

Install

Requires Python 3.11+ and macOS. See docs/install.md for the full prerequisites and Playwright notes.

pip install git+https://github.com/skoonin/daily-driver.git
daily-driver --version

Quick start

daily-driver init ~/daily-driver-workspace
cd ~/daily-driver-workspace
daily-driver doctor
daily-driver tracker add --category task --title "Write phase 7 tests"
daily-driver status

Full first-day walkthrough in docs/usage.md. Bare-minimum scaffold steps in docs/quick-start.md.

Job search quick start

The job-search plugin scrapes a configurable set of boards into jobs.csv, dedupes against what's already there, and enriches each row (fit, notes, comp) via claude or a local Ollama model.

1. Add a plugins.job_search block to .dd-config.yaml and turn the scraper on with scraper.enabled: true. Listing a source key enables it (disable with enabled: false):

plugins:
  job_search:
    persona: "Senior SRE, IC-track only"
    seniority_keywords: [senior, staff, principal]
    locations:
      remote: true
      # ISO code -> city allow-list; [] = whole country, cities = only those
      countries: {US: [], CA: [Vancouver, Victoria]}
    scraper:
      enabled: true
    sources:
      greenhouse: {enabled: true, greenhouse_boards: [company-a, company-b]}
      ashby:      {enabled: true}   # boards come from discover-boards below
      lever:      {enabled: true}   # ditto
      remoteok:   {remoteok_tags: [devops, kubernetes, aws]}
      hn_who_is_hiring: {}

2. Discover which Greenhouse / Ashby / Lever boards list your roles. This sweeps the ATS slug universe and caches the matches for every later jobs run — the first sweep probes everything and takes tens of minutes; later sweeps are incremental: new slugs, plus boards overdue for re-probe, up to discovery.max_reprobe_per_sweep per platform per sweep:

daily-driver jobs discover-boards

3. Run the search, then top up any cells a run left empty:

daily-driver jobs run           # scrape + enrich -> jobs.csv
daily-driver jobs backfill      # re-enrich empty cells from an earlier run
daily-driver jobs status        # last-run metadata + row counts

Full field reference in configuration.md; the complete command set, source list, and AI-provider choice in usage.md.

Workspace layout

<workspace>/
  .dd-config.yaml          # workspace marker + structured config
  context.md               # your profile (seeded once, edit freely)
  voice-profile.md         # writing voice (ditto)
  tracker.yaml             # tracker entries (CLI-managed)
  jobs.csv                 # scraper output (plugin-managed)
  .daily-driver/
    version                # version stamp
    manifest.json          # SHA-256 manifest of managed .claude/ files
    state/                 # ephemeral: locks, logs
  .claude/
    commands/daily-driver/ # shipped slash commands (regenerated on drift)
    agents/daily-driver/   # shipped agents (ditto)
    commands/*.md          # your custom commands (untouched)
    agents/*.md            # your custom agents (untouched)
  <YYYY>/<MM>/             # daily plans + notes

Workspace discovery walks up from CWD looking for .dd-config.yaml. Override with --workspace PATH.

Documentation

Full index: docs/README.md.

Users — start with quick-start.md to scaffold, then usage.md for the daily flow. Read concepts.md once for the mental model.

Developers — start with docs/dev/developer.md.

Doc	What it covers
concepts.md	Mental model: workspace layout, surfaces, design decisions
quick-start.md	Minimal scaffold checklist
usage.md	End-to-end daily flow with worked examples
commands.md	Every subcommand and non-obvious flags
configuration.md	.dd-config.yaml schema and customization
ollama-setup.md	Local LLM provider for enrichment/summary
troubleshooting.md	Failure modes and recovery
dev/developer.md	Architecture, runtime flow, init contract
dev/extending.md	Adding subcommands and scraper sources
dev/releasing.md	Release workflow, semver, CHANGELOG

See CONTRIBUTING.md for dev setup and commit conventions.

Platform

macOS only. pbcopy, icalBuddy, AppleScript, and launchd integrations assume macOS. Linux is on the roadmap.

License

MIT.