moonlighter MCP server
Commit Nov 14, 2026Scans job boards, scores postings against your profile, and drafts application answer sheets and tailored CVs.
Also known as albertosca/moonlighter
- Stars
- 11
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 14, 2026
- License
- AGPL-3.0
What is moonlighter MCP?
moonlighter scans job boards, scores postings against your profile, and drafts application answer sheets and tailored CVs.
Summary generated from the project's README
About moonlighter MCP server
Job seekers looking to automate their application workflow can use moonlighter to monitor job boards and prepare tailored materials. By connecting through an MCP client like Claude or using the Claude Code CLI, the tool scans specific companies on platforms such as Greenhouse, Lever, Ashby, Recruitee, Workable, SmartRecruiters, and InHire. It evaluates new postings against a user profile and assigns a score to determine relevance.
Once a suitable role is identified, the server drafts a complete answer sheet for application forms and can generate a one-page LaTeX CV tailored to the specific posting. To maintain user control, the system never submits applications automatically; instead, it provides the drafted content for manual review and pasting. All job history and drafts are stored locally in a SQLite file under MOONLIGHTER_HOME. Users need an Anthropic API key or the Claude Code CLI to power the underlying intelligence.
What you can do with it
- 1Score job postings for relevanceEvaluate how well a new job description matches your professional profile.
- 2Draft application form responsesGenerate answers for screening questions based on your existing profile data.
- 3Create tailored resumesProduce a one-page CV specifically adjusted for a particular job description.
Written from the project's README. Every use case is checked against a quote from it.
Automated checks
Ran Nov 14, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablealbertosca/moonlighter is public and reachablegithub.com/albertosca/moonlighter ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)AGPL-3.0spdx.org/licenses/AGPL-3.0.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-13github.com/albertosca/moonlighter ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubmoonlighter
The job-application assistant that leaves the last word to you.
Scans the job boards you choose, scores each posting against your profile, and drafts a complete answer sheet from your own data, flagging what it can't answer.
Built by Alberto Cavalcanti · Connect on LinkedIn · Read the docs · Install
Sixty seconds of it
You: scan my companies
moonlighter: 3 sources scanned — 41 postings, 38 already known, 3 new
✓ NEW — Acme Robotics / Senior Backend Engineer
Score: 8.1/10 (threshold: 6.5)
✓ NEW — Nimbus Health / Staff Engineer
Score: 7.4/10
✗ Vandelay Industries / .NET Architect — 3.2/10, archived (hard filter: .NET)
You: prepare the application for the Acme one
moonlighter: Senior Backend Engineer — Acme Robotics
[1/9] Full name (required)
Ana Lima
[2/9] Email (required)
ana.jobs+k3xv9qhm@gmail.com ← tracking alias: Acme's reply will match this application by itself
[3/9] Why do you want to work at Acme? (required)
Three sentences drafted from Ana's profile.
[4/9] Desired salary (required)
BRL 28.000/month
[5/9] Do you hold a US work visa? (required)
!! I DON'T KNOW — no basis in your profile to answer
1 of 9 need you
You: (review it, paste it into the form, hit send — moonlighter never does)
The conversation is illustrative: Claude relays the tool output in its own words. The gap is real behaviour — the model is told to answer UNKNOWN when your profile gives it no basis, and the sheet hands that question back to you.
How it works
- Scan — reads job boards on Greenhouse, Lever, Ashby, Recruitee, Workable, SmartRecruiters and InHire for the companies you list, plus optional job portals you switch on in config.
- Evaluate — an LLM scores each posting against your profile; postings below your threshold are archived.
- Prepare — drafts an answer for every question on the form into one sheet you review, paste and send; optionally, a one-page CV tailored to the posting.
- Track — matches recruiter replies in Gmail to the right application by its
+aliasand moves your pipeline forward.
You drive it from a Claude conversation through MCP tools; the scan, apply and email packages also install JSON command lines for shells and cron jobs.
What it won't do
- It never sends an application. It drafts the sheet; you paste the answers into the employer's form and send it yourself.
- Your pipeline lives in a local SQLite file. Jobs, drafts and application history are stored under
MOONLIGHTER_HOME. What leaves your machine is only what goes to the LLM and the APIs you configure — Claude, Gmail, the job boards. PRIVACY.md has the details. - The model never answers a question the guards recognise as salary, compliance or demographic. Those come from your profile or config, or are left to you. The guards match known phrasings, and when you paste in a page's text, the model still reads all of it to find the questions.
The model can still get an answer wrong — which is why every sheet is yours to review.
Engineering decisions
Applications go out under your name, so the bar is trust. Each decision below bought safety at a price; the Engineering page records what each one cost and where to check it in the code.
| Decision | Trade-off we accepted |
|---|---|
| Stopped driving the browser | Every application costs you a paste and a click |
| Deterministic guards around the drafting step | They recognise known phrasings only |
| Model text escaped into the CV's LaTeX, never validated | Model output can't carry LaTeX formatting beyond bold |
| Five slices with a tested import boundary and lockstep releases | Every release bumps five packages by hand |
| 100% branch coverage as a gate, gates proven by canaries | Every new branch costs a test |
Install
You need uv (it fetches Python 3.14 for you) and an LLM backend: the Claude Code CLI by default, or an Anthropic API key.
- Claude Code:
/plugin marketplace add albertosca/moonlighter, then/plugin install moonlighter@moonlighter. - Any other MCP client: register the server command
uvx moonlighter. - Then:
uvx moonlighter initwrites your config; fill inprofile.yamlandcompany_list.yaml, and ask Claude to "scan my companies".
Documentation
- Getting started — requirements, the setup wizard, registering the MCP server
- Tailored CV — a one-page LaTeX CV per posting, built from bullets you curated
- Answer bank — screening answers you approved, reused on the next application
- Command line — a JSON CLI per slice, exit codes, what each install gives you
- MCP tools — the 18 tools Claude calls on your behalf
- Extensions — add a job source as a separate package
Questions and troubleshooting: FAQ. Working on the code: CONTRIBUTING.md.
License
AGPL-3.0 — see LICENSE: use it, fork it, modify it, as long as what you distribute or serve over a network stays open. Want to offer moonlighter as a hosted or paid service without the AGPL's obligations? A commercial license is available — open an issue to start that conversation; the CLA every contributor signs keeps that offer possible.
DISCLAIMER.md covers terms of service, automation and LLM backend usage; PRIVACY.md covers what the tool stores and where it goes.
Built by Alberto Cavalcanti — Connect on LinkedIn
moonlighter MCP questions
Does moonlighter send job applications automatically?
No, the tool only drafts the answer sheets for you to review, paste, and send yourself.
Where is my job application data stored?
Your pipeline, including jobs, drafts, and history, is stored locally in a SQLite file under MOONLIGHTER_HOME.
How does the tool handle sensitive questions like salary or demographics?
The model is prevented from answering questions identified as salary, compliance, or demographic; these are instead pulled from your profile or left for you to answer.