GitEmployed

A serverless, GitOps-driven job application and tracking system. GitHub Issues are your pipeline, GitHub Projects is your Kanban board, and GitHub Actions is the automation plane that scrapes roles, AI-triages fit, and tailors your resume — all running free on GitHub's infrastructure.

Getting Started

1. Replace the placeholder resume in resumes/resume.yaml (a YAML file in JSON Resume format) with your real work history, then commit and push to main. The daily cron takes over from there.

Configuration

Per-user options live in config/settings.yaml and are edited directly in that file — same workflow as replacing the placeholder resume. Notable defaults:

• search.work_preference defaults to hybrid (also accepts remote or onsite)
• Your resume's basics.location fields (city, state, countryCode) drive the search location — include all three for best results with hybrid/onsite modes
• search.enabled (set to false to pause daily scraping)
• fit_threshold (minimum fit score, 1.0–5.0, default 3.5)

Resume theme

theme controls how your tailored resume is rendered to PDF. It accepts any theme from the JSON Resume theme ecosystem — browse npm or GitHub for options — in one of two pinned forms:

• An npm package pinned to an exact version: "<package>@<version>", e.g. "@jsonresume/jsonresume-theme-professional@1.0.22"
• A GitHub-only theme pinned to a full 40-character commit SHA: "github:<owner>/<repo>#<sha>"

GitEmployed installs the theme package — running its own code in the process — and later renders with it, so always pin to an exact version or commit SHA, never a floating branch, tag, or dist-tag like latest, since a theme maintainer could otherwise push new code that runs unreviewed in your repo's Actions. Omit this key to use the built-in default theme.

Docs

Full setup, configuration, and workflow documentation: https://github.com/menil/gitemployed

Your Job Search on GitHub

Issues

Each scraped job listing becomes a GitHub Issue. The AI triage engine scores every listing against your resume, tailors a fit resume for strong matches, and closes mismatches with a reasons comment.

Projects

An optional Kanban board for tracking your application lifecycle (Triage Pending → Ready to Apply → Applied → In Loop → Rejected). Requires Project V2 setup; without it, issue labels track state instead.

Secrets & Variables

Configure these under Settings > Secrets and variables > Actions:

Secrets

Secret	Purpose
GEMINI_API_KEY	Gemini LLM key for triage and tailoring (Google AI Studio) — one of Gemini, OpenRouter, or Claude is required
OPENROUTER_API_KEY	OpenRouter LLM key — also powers automated PR reviews
CLAUDE_CODE_OAUTH_TOKEN	Claude provider credentials: Claude Code subscription token (claude setup-token) or standard Anthropic API key
GH_PAT	Optional. GitHub Personal Access Token. Enables status badges via Gist, and optionally Projects V2 Kanban board automation; falls back to GITHUB_TOKEN without it
TAVILY_API_KEY	Optional. Enables the Tavily search provider for the Issue Assistant's web research
BRAVE_API_KEY	Optional. Enables the Brave search provider for the Issue Assistant's web research
JINA_API_KEY	Optional. Raises page-fetch rate limits from 20 to 500 RPM for JS-heavy job boards
GMAIL_CLIENT_ID	Optional. Required only to enable Gmail Sync — see DEVELOPMENT.md for the one-time OAuth setup
GMAIL_CLIENT_SECRET	Optional. Required only to enable Gmail Sync (paired with GMAIL_CLIENT_ID)
GMAIL_REFRESH_TOKEN	Optional. Required only to enable Gmail Sync; long-lived, read-only (gmail.readonly)

Variables (Optional)

Configure these under the Variables tab:

Variable	Default	Purpose
LLM_PROVIDER	auto-detected	gemini, openrouter, or claude (auto-detected from keys when unset)
OPENROUTER_MODEL	openrouter/free	OpenRouter model (requires provider prefix, e.g., openrouter/free or google/gemini-2.5-flash).
GEMINI_MODEL	models/gemini-2.5-flash	Gemini model (e.g., models/gemini-2.5-flash or gemini-2.5-flash)
CLAUDE_MODEL	claude-sonnet-5	Claude model (e.g., claude-sonnet-5 or claude-haiku-4-5-20251001)

Actions

Workflow	What it does
scrape-jobs.yml	Daily cron that discovers new roles from LinkedIn, Indeed, and ZipRecruiter, then opens issues
triage-issue.yml	AI-scores each listing against your resume and tailors a fit resume if it passes the threshold
respond-issue.yml	Answers questions on issue threads via live web research, and applies status labels from conversation
status-transition.yml	Moves your board card when you apply a lifecycle label (applied, in-loop, ...), and closes the issue for terminal labels (rejected, triage-mismatched) while keeping their columns distinct
project-status-sync.yml	Reverse sync: applies the matching label when you drag a card to a new column
gmail-sync.yml	Optional, off by default. Hourly cron that matches lifecycle emails in a label-scoped slice of your Gmail inbox to open applications and applies the same status updates as the Issue Assistant