Skip to main content

JobHunter MCP Server

PyPI version Python versions License: MIT Smithery

JobHunter gives an AI assistant (Claude Desktop, Cursor, Antigravity, VS Code) tools for a full job search. It searches many sources, scores each job against your real resume, writes application documents from facts in that resume, and tracks every application. All data stays on your machine.

JobHunter works like a small company. You talk to one AI, and that AI is the CEO. It briefs a team of specialists (Scout, Analyst, Writer, Coach, Chief of Staff). A Reviewer checks every job and document before you see it. See docs/team.md.

By default the AI sees seven CEO tools. The 39 specialist tools sit behind them. --surface full shows all 46. There are no MCP prompts. The old prompt contracts are the get_playbook tool.

Quick start

🚀 1-Command Run (Published on PyPI — Zero Clone, Zero Setup)

You do not need to clone this repository to run JobHunter with your AI assistant. Run it instantly with a single command via uvx (the fast, isolated package runner from Astral) or Smithery:

1. Instant Verification / Terminal Run

uvx jobhunter-mcp --check

2. Claude Desktop Integration

Add JobHunter to your claude_desktop_config.json:

{
  "mcpServers": {
    "jobhunter": {
      "command": "uvx",
      "args": ["jobhunter-mcp"]
    }
  }
}

3. Cursor Integration (.cursor/mcp.json or Settings > MCP)

{
  "mcpServers": {
    "jobhunter": {
      "command": "uvx",
      "args": ["jobhunter-mcp"]
    }
  }
}

4. Smithery Integration (1-Click)

npx -y @smithery/cli run @ErenYea9er69/jobhunter

Or 1-click install directly into Claude Desktop:

npx -y @smithery/cli install @ErenYea9er69/jobhunter --client claude

5. Optional Arguments

You can pass any arguments directly to uvx jobhunter-mcp:

# Run with all specialist tools exposed
uvx jobhunter-mcp --surface full

# Run with custom resume workspace directory
uvx jobhunter-mcp --workspace /path/to/my-resumes

🛠️ Developer Setup & Local Copilot Suite (Optional)

Clone only if you are developing JobHunter, customizing the codebase, or using the standalone local Web Dashboard and Chromium extension:

git clone https://github.com/ErenYea9er69/jobhunter.git
cd jobhunter
pip install -r requirements.txt

Run on Windows:

Double-click run.bat or execute in PowerShell / CMD:

run.bat
# or
python run.py

Run on macOS & Linux:

chmod +x run.sh
./run.sh
# or
python3 run.py

Launcher CLI Options:

python run.py [OPTIONS]
  --port PORT          Port for HTTP daemon (default: 8765)
  --host HOST          Host to bind to (default: 127.0.0.1)
  --no-browser         Do not auto-launch Chromium with extension
  --setup-only         Provision directories, DB, and MCP configs, then exit
  --set-key KEY        Store API key for Standalone Direct API mode
  --provider PROVIDER  AI Provider (gemini, openai, anthropic, deepseek, groq, mistral)
  --model MODEL        Target model (e.g. gemini-3.8-flash, gpt-5-mini, claude-3-7-sonnet)
  --clear-key          Delete stored API keys and return to MCP mode
  --ai-status          Display active AI provider, mode, and masked key
  --test-ai            Test connection to the active AI endpoint and measure latency

What happens automatically:

  1. Candidate Drop Folder Provisioned: Auto-creates resumes/ with a PUT_YOUR_RESUME_HERE.txt guide. Drop your Word (.docx), PDF (.pdf), or Text (.txt, .md) resume right into this folder.
  2. SQLite Database Initialized: Automatically creates your private local database (engine.db) in ~/.jobhunter/ or your workspace.
  3. AI Coding Agents Auto-Connected (MCP): Auto-detects and writes MCP configuration for Cursor (.cursor/mcp.json) and Claude Desktop so your AI assistant shares the exact same resume facts and tracking database.
  4. Web Dashboard Served: Starts the local HTTP daemon and serves the 6-tab autonomous dashboard on http://127.0.0.1:8765.
  5. Browser Extension Loaded: Detects your installed Chromium browser (Google Chrome, Microsoft Edge, Brave) and launches it with the ATS Autofill extension preloaded (--load-extension).
  6. First-Run Web Drop Wizard: If no resume exists yet, the web dashboard displays a drag-and-drop wizard allowing you to upload your Word, PDF, or Text resume directly from the browser with instant skill extraction.

🖥️ Modern Web Dashboard & Visual Workspace

JobHunter includes a responsive, dark/light themed single-page web application running locally on http://127.0.0.1:8765. The interface uses a black, gray and white theme with small green, amber and red score colors. See DESIGN.md and UI-CHANGES.md:

┌────────────────────────────────────────────────────────────────────────┐
│                          JOBHUNTER DASHBOARD                           │
├───────────────┬───────────────────────────────┬────────────────────────┤
│ TAB 1: RADAR  │ TAB 2: PIPELINE KANBAN        │ TAB 3: DOCUMENT STUDIO │
│ Opportunity   │ 5-Stage Visual Funnel         │ 100% Truthful QA       │
│ Discovery     │ Saved → Applied → Interview   │ Cover Letters & Pitch  │
├───────────────┼───────────────────────────────┼────────────────────────┤
│ TAB 4: PROFILE│ TAB 5: AI & CALIBRATION       │ TAB 6: AUTO-APPLY HUB  │
│ Resume Drop   │ Dual Engine (MCP vs API)      │ ATS Extension Dispatch │
│ & Variants    │ 2026 Frontier Models          │ 1-Click Browser Launch │
└───────────────┴───────────────────────────────┴────────────────────────┘

Tab 1: Radar & Opportunity Discovery

  • Multi-Source Aggregation: Real-time unified search across 100+ platforms (Company Greenhouse/Ashby/Lever boards, Y Combinator, Hacker News, regional national sites, and remote boards).
  • Match Scoring Engine (0-100%): Every posting is ranked strictly against the candidate's verified skills, experience level, and target preferences.
  • Smart Filters: Filter instantaneously by Work Mode (Remote, Hybrid, Onsite), Job Type (Full-time, Internship/PFE, Contract), and Score Floor.
  • 1-Click Queue Auto-Apply: Queue promising opportunities directly into the Auto-Apply batch runner.

Tab 2: Application Kanban Pipeline

  • Visual Funnel Tracking: 5 automated progression lanes: Saved → Queued Auto-Apply → Applied → Interviewing → Offers.
  • Gmail Sync Integration: 1-click sync scans your connected inbox for ATS confirmations, interview invitations, and status updates, automatically transitioning opportunities across lanes without manual entry.
  • Application Analytics: Tracks submission velocity, response rates, and active interview conversions.

Tab 3: Truthful Document Studio & Copilot

  • Multi-Format Document Generation:
    • Cover Letters: Direct-impact, metrics-driven pitches tailored to the job description.
    • YC-Style Founder Cold Emails: High-converting, short pitches with an automated Founder Corporate Email Finder that synthesizes and verifies executive email formats.
    • Big Tech Recruiter & Internal Referrals: Calibrated for levels and terminology at Google (L3-L6), Meta (E3-E6), Uber, Amazon, Apple, Microsoft, Netflix, and Stripe.
    • STAR Interview Guides: Complete preparation matrices with Situation, Task, Action, and Result outlines grounded in your actual work history.
  • Reviewer QA Badge: Every generated paragraph is audited sentence-by-sentence against candidate resume facts before approval. Unsupported claims are rejected or flagged with zero hallucinations.
  • Direct Export: 1-click Copy, Print to PDF, or "Open in Gmail" prefilled draft.

Tab 4: Candidate Profile & Resume Variants

  • Master Profile Display: Visual hero card showing extracted coordinates, verified skills chips, and reverse-chronological professional timeline.
  • First-Run Drop Folder & Drag-and-Drop Web Wizard: Upload Word (.docx), PDF (.pdf), or Text (.txt, .md) resumes directly in the browser or drop them into resumes/.
  • Variant Management: Switch between specialized resume variants (e.g. Backend vs AI Engineer) with one click; all ranking and generation immediately recalibrate.

Tab 5: Search Calibration & AI Dual Engine

  • Target Filtering: Configure target countries (e.g., Tunisia, France, Germany, Worldwide Remote), preferred cities, work mode, and search depth.
  • Dream & Excluded Companies: Boost priority for dream employers; permanently blacklist companies and unwanted technologies.
  • AI Engine Control Deck: Live mode badge (Direct API Active vs External MCP Host Mode), provider selector, 2026 frontier model picker, masked key display, latency test, and 1-click key wipe.

Tab 6: Auto-Apply & Desktop Hub

  • Queue Dispatcher: Automatically iterates through queued applications and launches the autofill sequence.
  • Work Authorization & Visa Compliance: Sets deterministic rules for legal authorization, sponsorship requirements, and notice period.
  • 1-Click Chromium Launcher: Detects your installed Chrome, Edge, or Brave browser and boots it with the JobHunter extension active.

🧩 Chrome ATS Autofill Extension

Located in extension/, the JobHunter extension turns your browser into an automated job application copilot:

  • Manifest V3 Architecture: Fast, lightweight, and operates entirely against your local daemon (http://127.0.0.1:8765).
  • Supported ATS Platforms:
    • Greenhouse (boards.greenhouse.io)
    • Lever (jobs.lever.co)
    • Ashby (jobs.ashbyhq.com)
    • Workday (*.myworkdayjobs.com)
    • SmartRecruiters, Recruitee, Workable, and custom employer career portals.
  • Smart Form Autofill: Instantly populates personal coordinates, resume attachments, LinkedIn, GitHub, portfolio URLs, and work authorization disclosures.
  • Grounded Screening Question Solver: Intercepts custom employer questions (e.g., "Describe a time you solved an architectural bottleneck", "What is your experience with Kubernetes?") and formulates accurate answers drawn solely from candidate facts in SQLite.
  • Review panel: A panel on the page lists every filled field with an Undo, lists fields that need you, and never submits unless you chose countdown mode.


🔀 Dual Execution Modes: External MCP Host vs. Standalone Direct API

JobHunter operates seamlessly in two distinct modes, and you can switch between them at any time:

┌────────────────────────────────────────────────────────────────────────┐
│                        MODE 1: EXTERNAL MCP HOST                       │
│  Zero API keys required in JobHunter!                                  │
│  Claude Desktop or Cursor runs its own frontier LLM model              │
│  and talks to JobHunter via MCP stdio tools.                           │
└────────────────────────────────────┬───────────────────────────────────┘
                                     │ (Switch dynamically anytime)
┌────────────────────────────────────▼───────────────────────────────────┐
│                    MODE 2: STANDALONE DIRECT API                       │
│  No external AI IDE needed! Clone and supply your own API key.        │
│  Powers the Web Dashboard & Chrome Extension directly via SQLite.      │
│  Supports Gemini 3.8 Flash, GPT-5 Mini, Claude 3.7 Sonnet, etc.        │
└────────────────────────────────────────────────────────────────────────┘

Mode 1: External MCP Host Mode (Zero Keys Needed)

  • If you use Claude Desktop, Cursor, Antigravity, or Claude Code, you do NOT need to supply any API keys.
  • The AI IDE hosts the conversation and passes context to JobHunter's tools over stdio.
  • All documents, tailoring, and scoring run through JobHunter's strict offline Reviewer QA engine.

Mode 2: Standalone Direct API Mode (Web Dashboard & Chrome Extension)

  • If you want to use the Web Dashboard and ATS Autofill extension independently without running Claude Desktop or Cursor, you can supply your API key directly via CLI or Web UI.
  • Your key is stored securely in your private local SQLite database (engine.db) and is never sent to external telemetry or exposed in logs.
Configure via Command Line (Windows, macOS, Linux):
# 1. Activate Google Gemini (Default frontier: gemini-3.8-flash)
python run.py --set-key YOUR_GEMINI_API_KEY --provider gemini --model gemini-3.8-flash

# 2. Activate OpenAI (Default frontier: gpt-5-mini or o3-mini)
python run.py --set-key YOUR_OPENAI_API_KEY --provider openai --model gpt-5-mini

# 3. Activate Anthropic Claude (Default frontier: claude-3-7-sonnet)
python run.py --set-key YOUR_ANTHROPIC_API_KEY --provider anthropic --model claude-3-7-sonnet

# 4. Activate DeepSeek
python run.py --set-key YOUR_DEEPSEEK_KEY --provider deepseek --model deepseek-chat

# 5. Check active mode & masked key
python run.py --ai-status

# 6. Test connection latency and endpoint health
python run.py --test-ai
Configure via Web Dashboard:
  1. Open http://127.0.0.1:8765 in your browser.
  2. Navigate to the Calibration & Filters tab.
  3. In the AI Engine & Dual Execution Mode card:
    • Select your provider (Google Gemini, OpenAI, Anthropic, DeepSeek, Groq, Mistral, or local Ollama).
    • Pick a model or enter a custom identifier.
    • Paste your key and click Save & Activate.
    • Click Test Connection to verify round-trip ping and latency.

Reverting to MCP Mode / Deleting Your Key:

You can wipe your stored API key at any time to return JobHunter to External MCP / Deterministic mode:

# Via CLI:
python run.py --clear-key

# Or via Web Dashboard:
Click the "Clear Key" button in Tab 5 (Calibration & Filters)

2026 Frontier Models Supported:

  • Google Gemini: gemini-3.8-flash (recommended default), gemini-3.1-pro, gemini-3.5-flash
  • OpenAI: gpt-5-mini (recommended default), o3-mini, gpt-5
  • Anthropic: claude-3-7-sonnet (recommended default), claude-3-5-haiku
  • DeepSeek: deepseek-chat, deepseek-reasoner
  • Groq: llama-3.3-70b-versatile, qwen-2.5-coder-32b
  • Mistral: mistral-large-latest
  • Ollama: llama3.3, deepseek-r1 (local, zero cloud keys)


Alternative: Standalone MCP Server via uvx

If you only want the stdio MCP server without the web dashboard:

uvx --from git+https://github.com/ErenYea9er69/jobhunter jobhunter-mcp

Claude Code CLI

Add JobHunter directly with the Claude CLI:

claude mcp add jobhunter -- uvx --from git+https://github.com/ErenYea9er69/jobhunter jobhunter-mcp

Claude Desktop Configuration

Add the server to your claude_desktop_config.json:

  • Windows: %APPDATA%\Claude\claude_desktop_config.json
  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "jobhunter": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/ErenYea9er69/jobhunter",
        "jobhunter-mcp"
      ],
      "env": {
        "JH_DATA_DIR": ""
      }
    }
  }
}

(Leave JH_DATA_DIR empty to default to ~/.jobhunter, or provide a custom directory path).

Local Development

pip install -r requirements.txt
python mcp_server.py --check        # starts the server, lists the tools, exits

Put a PDF, DOCX or TXT resume in your workspace folder (or in resumes/), or use scan_resumes / parse_resume_file. Local configuration can also be pointed at a workspace:

{
  "mcpServers": {
    "jobhunter": {
      "command": "python",
      "args": ["/absolute/path/to/jobhunter/mcp_server.py"],
      "env": { "JH_WORKSPACE": "/absolute/path/to/your/workspace" }
    }
  }
}

First conversation. Just talk to the AI:

  1. "Set me up. I want to work in Tunisia and France." The CEO reads your resumes and saves your preferences.
  2. "Find backend engineer jobs." The Scout searches, the Reviewer checks each job and reads the top posting pages.
  3. "Find small companies hiring python developers in Sfax." Same, with a bias toward small and local employers.
  4. "Write a cover letter for the second one." or "...for the Adyen job." The CEO resolves the pointer against the last list. The Writer drafts. The Reviewer checks every sentence against your resume.
  5. "I applied to Adyen." "Not interested in banks." "Why is the fit low for the first one?" The CEO updates your tracker or your exclusions, then reads the change back to confirm it. Or it explains the score.

The answer is a short brief. The full report and data are behind it. Every answer says what was not verified. Nothing is submitted for you.

Make it yours

set_preferences stores your choices in .jobhunter/preferences.json. They apply to every resume and every search.

Key Meaning
target_countries Where you want to work, for example ["Tunisia", "France", "Germany"]. Picks the local sources.
home_country, cities Where you live and the cities you prefer.
work_mode Any, Remote, Hybrid or Onsite. Onsite skips remote-only boards.
target_titles Roles to search. The first is the main one.
min_salary, salary_currency A yearly floor. Salaries in TND, MAD, EGP, AED and SAR are read as monthly figures.
job_types, seniority Full-time, Contract, Internship. Junior to director.
excluded_keywords, excluded_companies, dream_companies Hide or boost.
company_focus any, avoid_big_names or only_big_names.
max_age_days Hide postings older than this. Default 30.
search_depth quick, standard or deep (more query variants).
enabled_sources, disabled_sources Choose exactly which sources run.
needs_visa_sponsorship, open_to_relocation, languages Stored and applied to the resume profile.

Example: set_preferences({"target_countries": ["Tunisia", "France"], "company_focus": "avoid_big_names", "excluded_keywords": ["php"]}).

Where the jobs come from

Which sources run depends on your target countries. A search from Tunisia no longer calls India's Naukri or France's HelloWork. A search that targets France calls HelloWork and fr.indeed.com.

Source Serves Notes
Company Boards Worldwide Greenhouse, Lever (US and EU), Ashby, Workable, SmartRecruiters, Recruitee, Personio, Teamtailor, BambooHR, Breezy
Y Combinator, Hacker News Worldwide Hacker News now reads three pages of the monthly thread
LinkedIn Worldwide Unchanged. See "Protected code" below
Indeed 41 countries One call per target country, on the local site
Bayt Middle East and North Africa One call per target country
Arbeitsagentur Germany Public employment service, no key
HelloWork, Naukri, ZipRecruiter France, India, US and Canada
Adzuna 19 countries Needs a free key (ADZUNA_APP_ID, ADZUNA_APP_KEY)
Regional Boards Tunisia, Morocco, Algeria, Egypt, Nigeria, Kenya, South Africa Local boards (Keejob, ReKrute, Wuzzuf and others) read through their job data and sitemaps. Tries local-language queries too
Jooble Many countries Aggregator that collects small local boards. Needs a free key (JOOBLE_API_KEY)
JobTech Sweden Public employment service, no key
Himalayas, Remotive, Jobicy, RemoteOK, We Work Remotely, Working Nomads Remote Skipped when you want onsite work
Arbeitnow, The Muse Europe, US and UK
Custom Feed, Career Page, Job Sites Anywhere you point them Added with add_job_source

Reaching small and regional employers

The big boards show the same famous names. These features reach the rest:

  • add_job_source(kind="company", value="Acme") finds the company's public board by name, or reads a careers URL directly. Every open role is searched.
  • add_job_source(kind="feed", value=<rss url>) adds any RSS, Atom or JSON feed, such as a national job board. The feed is test-read before it is saved.
  • add_job_source(kind="career_page", value=<url>) reads schema.org job data on a careers page, or follows an embedded ATS link.
  • add_job_source(kind="site", value=<url>) adds a small job board or any site that publishes schema.org job data (what Google for Jobs reads). The reader finds job pages through the sitemap and the page's links. It obeys robots.txt and refuses a site that forbids reading, even with force. Not sure which kind to use? discover_job_source(url) explains how a site can be read and adds nothing.
  • Regional Boards runs the catalog in jh/regional_boards.py for your target countries. Queries are also sent in French for countries where local boards post in French ("software engineer" also searches "ingénieur logiciel").
  • Pinpoint is a new company board type, so small employers on that system are found by name.
  • Boards discovered along the way: every job link that points at a known ATS adds that company's board to later searches. Each search teaches the next one. Boards that fail three times rest for 24 hours.

The built-in starter list of company boards is short and its slugs are unverified. Dead ones prune themselves.

Honest limits of this reach: the catalog names real local boards, but nobody has checked here which of them publish readable job data or allow reading. A board that does not shows zero in source_health. Add your own with kind="site". A site that builds its job list with scripts cannot be read this way. Use kind="company" if the employer uses a known ATS.

Tools

The full list with arguments is in docs/tool-contracts.md. The CEO tools (ceo_run and six more) are described in docs/team.md. Specialist groups: resumes (5), preferences and sources (7), finding jobs (4), analysis (5), documents (7), interview and outreach (4), playbooks (1), tracking (5), server (1).

Every tool returns one envelope: status (ok, partial, error), success, data, warnings, sources, correlation_id. Errors carry a code and a hint. Search results are compact cards, so a search never floods the model's context. Use the job_id from a card with any tool.

Truthful by design

  • The Reviewer does not trust the employee who made the work. It checks every job (fields, link, date, country, duplicates, matched skills, scams, text aimed at your AI), reads live posting pages, and checks every document sentence against your resume. Invented employers, titles, degrees, certifications, skills, numbers and awards are removed or rejected. Details in docs/team.md.
  • The Reviewer is measured. python -m jh.team.evals scores it on fixed cases, and docs/evals.md reports the numbers and how far to trust them. The honest estimate on text it has never seen is lower than the scores on the cases it was built with.
  • The CEO's report comes from tool results only, and every job and document names the call that produced it.
  • The prompts follow their own rules. A linter checks every prompt for dashes, capital-letter shouting, persona boasting and invented statistics, and a test checks that each playbook asks the model for the keys the code reads. Drafts are scanned for machine-sounding phrases, and the Reviewer warns about them. See docs/prompts.md.
  • When a search finds nothing, the CEO widens it in steps and then says so. It does not fill the gap.
  • Generated text uses only facts the resume supports. get_evidence lists them.
  • verify_draft checks any text against that evidence and flags unsupported claims. tailor_resume, cover_letter and prepare_application run the same check and return the result.
  • Tools never accept or return API keys. Optional LLM keys come from environment variables (ANTHROPIC_API_KEY, OPENAI_API_KEY, GEMINI_API_KEY and others). Without a key the deterministic engine writes the documents.

Your data

Everything lives in your user home directory at ~/.jobhunter/ by default, or the folder configured via the JH_DATA_DIR environment variable.

  • engine.db is the SQLite database: resumes, stored jobs, applications, watchlist, board health.
  • preferences.json holds your preferences, feeds and career pages.
  • output/ holds adapted resumes, cover letters, interview prep and exports.

Applications live in SQLite. Any old applications.json is imported once and renamed.

Protected code

The LinkedIn scraper (LinkedInGuest), the LinkedIn radar in jh/pipeline.py, and its two config getters are not to be edited. tests/test_protected_linkedin.py checks them byte for byte. The radar's background workers are not started by the MCP server. See AGENTS.md.

Tests

pip install -r requirements-dev.txt
python -m pytest tests -q

The suite runs offline. A fake network serves fixtures that pin each vendor's response format. Live behaviour of the newer adapters (Recruitee, Personio, Teamtailor, BambooHR, Breezy, Working Nomads, The Muse, Arbeitsagentur, Jooble, JobTech, Pinpoint) and of the regional boards was not checked against real servers. The site reader and the Reviewer's link check follow public standards (schema.org JobPosting, sitemaps, robots.txt) and are tested against fixtures. source_health shows any adapter whose format has drifted.

Known limits

  • Job IDs ignore location, so one title at one company in two cities shares an ID. The LinkedIn radar uses the same key, so it was left alone.
  • Salary comparison does not convert currencies.
  • The CEO's request router is rule-based. It reads English and French, and a few Arabic words. On phrasings it had never seen, it was right about two times in three. An unclear request makes the CEO ask. The AI should pass hints.intent and hints.reference so nothing is guessed.
  • The Reviewer is also rules, not a model. It was built to catch what language models tend to invent. A new kind of fabrication can get through. Facts about the employer beyond the posting cannot be checked at all, so they are flagged.
  • When no model writes the text, the code's own templates do, and some still contain stock phrases (the outreach, follow-up and screening templates). The Reviewer flags them. They are listed in docs/prompts.md.
  • Only the top five jobs of a search get a live link check. Ask the CEO to recheck one before you apply.
  • Several scrapers return an empty list instead of an error when blocked, so source_health shows them as ok with zero results.
  • Reading some boards (LinkedIn, Indeed, ZipRecruiter and others) uses browser-like requests. See docs/data-policy.md and use disabled_sources if you do not want that.

License

MIT. See LICENSE.

Metadata

Release files for jobhunter-mcp 1.0.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for jobhunter-mcp 1.0.0
File Size Uploaded Provenance Actions
jobhunter_mcp-1.0.0.tar.gz 549.2 kB With provenance Details for jobhunter_mcp-1.0.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for jobhunter-mcp 1.0.0
File Interpreter ABI Platform Actions
jobhunter_mcp-1.0.0-py3-none-any.whl Python 3 none any Details for jobhunter_mcp-1.0.0-py3-none-any.whl

Total release size: 1.1 MB

Release files / jobhunter_mcp-1.0.0.tar.gz

Download URL jobhunter_mcp-1.0.0.tar.gz
Size 549.2 kB
Tags Source
SHA-256 checksum
How to use checksums
8f0417983adf2e9a8d6ce51099d539cbd69ff1da0df07249e41b3ef510de8bb9
BLAKE2b-256 checksum
How to use checksums
5f3dd4b5e226d5484f44eb0d37ab088cd807c3f4cb0fd4c95564b5cd0d743c74
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 16, 2026.

Transparency log

Release files / jobhunter_mcp-1.0.0-py3-none-any.whl

Download URL jobhunter_mcp-1.0.0-py3-none-any.whl
Size 504.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
b13ef7830475e60434988038b08b45bb4fd2a4009ed963fb44704098bfcc90fd
BLAKE2b-256 checksum
How to use checksums
0a59cb86db23b354a5678bacbb4d19dfa048507a733a0efbe62c902be888fc56
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 16, 2026.

Transparency log

Security for jobhunter-mcp 1.0.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for jobhunter_mcp-1.0.0.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 16, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

1.0.0 This release

2 release files Full provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page