ankush1699

Autonomous Job Agent — Development skill for Claude Code

Development community

Job-application agent: scrapes LinkedIn, Indeed, Dice, Google Jobs and company career pages, scores each posting against your profile with Claude Haiku (~$0.003/posting), and tailors a resume and cove.

How to install Autonomous Job Agent

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ankush1699/autonomous-job-agent and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Autonomous Job Agent does

Job-application agent: scrapes LinkedIn, Indeed, Dice, Google Jobs and company career pages, scores each posting against your profile with Claude Haiku (~$0.003/posting), and tailors a resume and cover letter with a 6-node LangGraph pipeline. Every generated claim is checked against your own data.

Alternatives in Development

  • Career-Ops For Codex — Read CLAUDE.md for all project instructions, routing, and behavioral rules 32.5k ★
  • Crypto Research Haiku Command — Execute lightweight cryptocurrency research by calling only haiku versions of crypto analysis agents in parall 3.4k ★
  • Resumeskills — 20 specialized skills for resume optimization, ATS analysis, interview prep, and career transitions 160 ★

README

AI Resume Agent

**Built for one candidate, then generalized.** The pipeline is data-driven (everything it may claim comes from your own `Ankush_Master_Data.json`), but a few prompts still describe the maintainer's profile (visa status, stack, seniority) — see *Using your own data* below. Personal project; not affiliated with any job board or employer.

An autonomous job-search pipeline: score a job posting for fit before spending any money on it, then tailor a resume and cover letter only for postings worth applying to. Three entry points (web UI, CLI, scraper) share one core engine.

Project structure

./
├── core/                   Provider-agnostic engine internals — the parts every
│   ├── llm.py                entry mode shares.
│   ├── red_flags.py         llm.py:        cheap-tier model resolution (Groq → Ollama →
│   ├── jd_cache.py                         Claude Haiku fallback) vs. quality-tier (Sonnet).
│   ├── should_apply.py      red_flags.py:  free, deterministic visa/clearance rejection regexes.
│   └── profile_cache.py     jd_cache.py:   permanent, content-hash-keyed cache — a posting
│                                           is never rescored twice, from any entry mode.
│                            should_apply.py: the 3-stage fit gate (red flags → cache → LLM
│                                              rubric) that everything else depends on.
│                            profile_cache.py: compact candidate-profile text for prompts.
│
├── engine/                 The LangGraph tailoring pipeline itself.
│   ├── graph.py              graph.py:      the 6-node graph (should_apply → triage →
│   ├── schemas.py                           strategist → writer → editor → finalizer)
│   ├── state.py                             plus the Python validators that check every
│   ├── pdf_generator.py                     generated claim against master data.
│   ├── generate_generic.py   pdf_generator.py: Jinja2 → LaTeX → PDF, plus the single
│