Skip to content
Back to search

Resume-Builder

AI-powered resume & cover letter generator with dual ATS + HR scoring. Works for any profession. Claude Code plugin.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/jananthan30--resume-builder/badge.svg)](/listings/jananthan30--resume-builder)

HTML

<a href="/listings/jananthan30--resume-builder"><img src="/api/listings/jananthan30--resume-builder/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Works with **Claude Code** (CLI/IDE), **Codex** (CLI/app/IDE), and **claude.ai** (web/Projects).

**Step 1: Install the plugin**

Claude Code:

/plugin marketplace add jananthan30/Resume-Builder
/plugin install resume-builder

Codex from a local checkout:

codex plugin marketplace add .

Then restart Codex and install **Resume Builder** from the **Resume Builder Local** marketplace.

**Step 2: Configure the runtime**

Claude Code exposes the plugin setup command:

/resume-builder:setup

This walks you through everything: - Checks if Python is installed (tells you where to download it if not) - Installs all dependencies automatically (`pip install -r requirements.txt`) - Creates your `config.json` with your name, email, phone, LinkedIn - Optionally links a Pro account for unlimited cloud scoring - Optionally sets up the LLM scorer (Claude API key)

For Codex, install Python 3.10+, run `python -m pip install -r requirements.txt`, and create `config.json` with a valid `master_resume_path`. The installed Codex surface exposes the Resume Team as `$resume-team`; it does not expose the Claude-style `/resume-builder:*` command namespace.

**Step 3: Start building resumes**

Claude Code:

/resume-builder:resume [paste a job description here]

Codex:

$resume-team [paste a job description here]

`$resume-team` publishes an authorized, digest-verified `resume.md` draft. It does not by itself create a DOCX or complete an application package.

Or find jobs first:

/resume-builder:find-jobs Senior Data Scientist in New York

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 16 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "resume-builder-marketplace",
  3. "description": "Native resume tailoring, auditing, scoring, and document-generation workflows.",
  4. "owner": {
  5. "name": "Resume Builder Contributors",
  6. "email": "support@getresumehq.com"
  7. },
  8. "plugins": [
  9. {
  10. "name": "resume-builder",
  11. "source": "./",
  12. "description": "AI-powered resume & cover letter generator with dual ATS + HR scoring. Works for any profession \u2014 software engineers, marketers, lawyers, finance, healthcare, and more.",
  13. "version": "1.2.0"
  14. }
  15. ]
  16. }
README.md · README · 921 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # ResumeHQ — the resume tool that refuses to lie for you
  2. Every AI resume tool promises to "beat the ATS." This one has a harder rule:
  3. **it never invents experience** — and when a job is a genuine mismatch, it
  4. declines to tailor at all and tells you why. Finds jobs, scores your fit with
  5. deterministic engines (not LLM vibes), tailors through a fail-closed pipeline
  6. where an independent auditor can veto the writer, and
  7. [publishes its own scoring failures](BENCHMARKS.md). Works as a
  8. [Claude Code](https://docs.anthropic.com/en/docs/claude-code) plugin, Codex
  9. plugin, or standalone web app.
  10. [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
  11. [![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
  12. [![Claude Code](https://img.shields.io/badge/Claude%20Code-Plugin-blueviolet)](https://docs.anthropic.com/en/docs/claude-code)
  13. [![Codex](https://img.shields.io/badge/Codex-Plugin-111111)](https://developers.openai.com/codex/plugins/build)
  14. ![Upload a resume, paste a job description, get match scores and recruiter-style feedback with fixes in seconds](assets/resumehq-demo.gif)
  15. *Upload a resume → paste the job posting → keyword-match and recruiter-style heuristic scores with concrete fixes, in under 30 seconds. [Try it free](https://getresumehq.com).*
  16. ---
  17. ## Why ResumeHQ?
  18. Most resume tools only score the resume you bring to them. ResumeHQ goes further:
  19. | Feature | Jobscan | Rezi | Teal | **ResumeHQ** |
  20. |---------|---------|------|------|--------------|
  21. | Keyword + semantic match scoring | ✅ | ✅ | ✅ | ✅ |
  22. | Recruiter-style heuristic review | ❌ | ❌ | ❌ | ✅ |
  23. | Discover matching jobs | ❌ | ❌ | ❌ | ✅ |
  24. | Score jobs against your resume | ❌ | ❌ | ❌ | ✅ |
  25. | Auto-tailor resume to JD | ✅ | ✅ | ❌ | ✅ |
  26. | ATS-compliant DOCX output | ❌ | ✅ | ❌ | ✅ |
  27. | Application tracker | ❌ | ❌ | ✅ | ✅ |
  28. | Works in Claude Code / claude.ai / Codex | ❌ | ❌ | ❌ | ✅ |
  29. | Open source | ❌ | ❌ | ❌ | ✅ |
  30. | [Publishes its scoring benchmarks](BENCHMARKS.md) — including the bugs | ❌ | ❌ | ❌ | ✅ |
  31. | **Refuses to fabricate** — declines to tailor genuine mismatches, audits every claim against your real resume | ❌ | ❌ | ❌ | ✅ |
  32. ---
  33. ## What This Does
  34. You paste a job description (or search for jobs). The system:
  35. 1. **Discovers** matching jobs from live job boards — scored and ranked by fit with your resume
  36. 2. **Gates candidate fit** — your master resume must clear the fit bar
  37. (default 50) against the exact JD with zero hard knockouts before resume
  38. work begins. If you don't fit, it says so and stops — no tool that invents
  39. experience to close the gap is working for you
  40. 3. **Analyzes** passing JDs — extracts keywords, required skills, domain, seniority level
  41. 4. **Tailors** your master resume — rewrites bullets, reorders sections, matches terminology
  42. 5. **Scores** the result with two independent advisory heuristics (keyword/semantic match + recruiter-style review)
  43. 6. **Iterates** automatically until the evidence match covers every must-have (75%+ evidence match) — an internal editorial target, not a prediction of any employer's screening
  44. 7. **Generates** production-ready DOCX files (resume + cover letter)
  45. 8. **Tracks** every application in an Excel spreadsheet
  46. The candidate-fit gate always runs first and cannot be bypassed by ATS/HR scores.
  47. After it passes, safe read/scoring work may run concurrently while authorization,
  48. DOCX generation, and tracker mutation remain ordered.
  49. ---
  50. ## Quick Start — 3 Steps
  51. Works with **Claude Code** (CLI/IDE), **Codex** (CLI/app/IDE), and **claude.ai** (web/Projects).
  52. **Step 1: Install the plugin**
  53. Claude Code:
  54. ```bash
  55. /plugin marketplace add jananthan30/Resume-Builder
  56. /plugin install resume-builder
  57. ```
  58. Codex from a local checkout:
  59. ```bash
  60. codex plugin marketplace add .
  61. ```
  62. Then restart Codex and install **Resume Builder** from the **Resume Builder Local** marketplace.
  63. **Step 2: Configure the runtime**
  64. Claude Code exposes the plugin setup command:
  65. ```
  66. /resume-builder:setup
  67. ```
  68. This walks you through everything:
  69. - Checks if Python is installed (tells you where to download it if not)
  70. - Installs all dependencies automatically (`pip install -r requirements.txt`)
  71. - Creates your `config.json` with your name, email, phone, LinkedIn
  72. - Optionally links a Pro account for unlimited cloud scoring
  73. - Optionally sets up the LLM scorer (Claude API key)
  74. For Codex, install Python 3.10+, run `python -m pip install -r requirements.txt`,
  75. and create `config.json` with a valid `master_resume_path`. The installed Codex
  76. surface exposes the Resume Team as `$resume-team`; it does not expose the
  77. Claude-style `/resume-builder:*` command namespace.
  78. **Step 3: Start building resumes**
  79. Claude Code:
  80. ```
  81. /resume-builder:resume [paste a job description here]
  82. ```
  83. Codex:
  84. ```
  85. $resume-team [paste a job description here]
  86. ```
  87. `$resume-team` publishes an authorized, digest-verified `resume.md` draft. It does
  88. not by itself create a DOCX or complete an application package.
  89. Or find jobs first:
  90. ```
  91. /resume-builder:find-jobs Senior Data Scientist in New York
  92. ```
  93. ---
  94. ## Claude Code Slash Commands (9)
  95. | Command | What It Does |
  96. |---------|-------------|
  97. | `/resume-builder:setup` | One-time setup wizard (installs Python deps, creates config, links Pro account) |
  98. | `/resume-builder:job-fit [JD]` | Deterministic master-vs-JD gate (fit bar, default 50, and zero hard knockouts) before tailoring |
  99. | `/resume-builder:resume [JD]` | Full application: tailored resume + cover letter + scoring + DOCX + tracking |
  100. | `/resume-builder:tailor-resume [JD]` | Resume only (no cover letter) |
  101. | `/resume-builder:cover-letter [JD]` | Cover letter only |
  102. | `/resume-builder:find-jobs [title] [location]` | Discover and score matching jobs from live job boards |
  103. | `/resume-builder:batch-resume` | Process multiple job descriptions in parallel |
  104. | `/resume-builder:writing-coach [file]` | Audit and rewrite resume bullets using 10 writing rules |
  105. | `/resume-builder:resume-team [JD]` | Publish an authorized `resume.md` draft through the native Researcher → Writer → Auditor → Editor workflow |
  106. If running Claude Code locally from the cloned repo, use short names: `/resume`,
  107. `/tailor-resume`, `/find-jobs`, etc. In Codex, invoke `$resume-team`.
  108. ### What Works Without Setup
  109. Some Claude Code commands can provide prompt-only previews before setup.
  110. Production resume generation through `/resume-builder:resume`,
  111. `/resume-builder:resume-team`, or Codex `$resume-team` requires Python,
  112. `config.json`, the deterministic candidate-fit preflight, and the evidence,
  113. human-voice, and canonical-integrity audit helpers; those gates are never skipped.
  114. | Command | Works immediately? | With setup? |
  115. |---------|-------------------|-------------|
  116. | `/resume-builder:job-fit` | No — requires the configured master and deterministic preflight | Digest-bound score, threshold, and hard-knockout decision |
  117. | `/resume-builder:resume` | No — the native team and deterministic audits require setup | Full audited resume + automated ATS/HR scoring and DOCX output |
  118. | `/resume-builder:resume-team` / `$resume-team` | No — requires macOS/Linux, the configured master resume, and Python audit helpers | Authorized, digest-verified `resume.md` draft; DOCX/tracker finalization is still pending |
  119. | `/resume-builder:cover-letter` | Yes — the assistant writes the letter | + DOCX output |
  120. | `/resume-builder:writing-coach` | Yes — full writing audit | Same |
  121. | `/resume-builder:find-jobs` | Yes — shows results (no score) | + ATS/HR fit scoring per job |
  122. | `/resume-builder:setup` | Yes — runs the setup wizard | N/A |
  123. | MCP scoring tools | No — needs Python | `evidence_match`, `score_resume`, `score_ats`, `score_hr`, `score_with_llm`, `explain_score`, `extract_text`, `discover_jobs` |
  124. ---
  125. ## MCP Tools (7 production-supported surfaces)
  126. After running `/resume-builder:setup`, the MCP scorer auto-starts and provides these tools that Claude Code or Codex can call natively:
  127. | Tool | What It Does |
  128. |------|-------------|
  129. | `evidence_match` | Requirement-by-requirement evidence matching, with the exact excerpt behind each conclusion (recommended) |
  130. | `score_resume` | Legacy ATS + HR analysis in one call |
  131. | `score_ats` | Keyword + semantic match scoring (8 heuristic components) |
  132. | `score_hr` | Recruiter-style heuristic review (6 factors + F-pattern) |
  133. | `score_with_llm` | LLM-augmented rubric scoring (requires ANTHROPIC_API_KEY) |
  134. | `explain_score` | Actionable improvement suggestions with missing keywords |
  135. | `extract_text` | Extract text from DOCX/PDF/MD/TXT files |
  136. | `discover_jobs` | Search live job boards and score each job against your resume |
  137. All listed MCP tools support **cloud-first scoring** — they try the cloud API first and fall back to local scoring automatically. Legacy direct rewrite endpoints or functions are not production-authorized tailoring paths. The capability-isolated native Resume Team is the sole production rewrite and draft-publication path.
  138. ---
  139. ## Job Discovery
  140. The `/find-jobs` command and `discover_jobs` MCP tool search live job boards and rank results by how well each job matches your resume — answering "which jobs should I actually apply to?" with data.
  141. ```
  142. /resume-builder:find-jobs Senior Product Manager in San Francisco
  143. /resume-builder:find-jobs Data Scientist remote
  144. ```
  145. **How it works:**
  146. 1. Searches Adzuna (16 countries, salary data) + Remotive (remote jobs) + JSearch (aggregated boards incl. niche career centers; optional RapidAPI key)
  147. 2. Pre-filters top 20 results by title relevance
  148. 3. Lightweight scores all 20 candidates (keyword + phrase + BM25 — fast)
  149. 4. Full ATS + HR scores top 10 finalists
  150. 5. Returns ranked list with scores, salary range, and apply links
  151. **Sample output:**
  152. ```
  153. Rank Title Company ATS HR Salary
  154. ──── ─────────────────────────── ───────────── ──── ──── ──────────────
  155. #1 Senior Data Scientist Pfizer 82% 74% $120k–$150k
  156. #2 Data Scientist II Goldman Sachs 79% 71% $110k–$140k
  157. #3 ML Engineer – NLP Microsoft 74% 68% $130k–$160k
  158. ```
  159. **API keys required for job search:**
  160. - **Adzuna** (free): Register at [developer.adzuna.com](https://developer.adzuna.com/) — add `ADZUNA_APP_ID` and `ADZUNA_APP_KEY` to your `.env`
  161. - **Remotive**: No key needed (remote jobs only, included automatically)
  162. ---
  163. ## Evidence Matching
  164. A universal "ATS score" does not exist. Recruiting systems parse, search,
  165. filter, evaluate requirements and sometimes rank — using materially different
  166. mechanisms — so a single percentage cannot describe them all.
  167. This tool answers a question that *is* answerable:
  168. > Your resume provides strong evidence for 17 of 21 important requirements.
  169. > One must-have has no evidence found. Eligibility is UNVERIFIED.
  170. Every conclusion cites the exact resume excerpt behind it, with character
  171. offsets you can verify.
  172. ```bash
  173. python evidence_match.py --resume resume.docx --jd job.txt --verify
  174. ```
  175. Three results stay separate and are never multiplied together:
  176. | Output | Meaning |
  177. |---|---|
  178. | **Eligibility** | PASS / FAIL / UNVERIFIED on explicit hard requirements |
  179. | **Qualification Evidence Fit** | How strongly the resume supports the role's requirements |
  180. | **Evidence Quality** | How explicit and traceable that evidence is |
  181. A missing licence is **UNVERIFIED**, not FAIL — the resume did not establish it,
  182. which is not proof the candidate lacks it. Only explicit contradictory evidence
  183. (an expired licence) produces FAIL.
  184. **Why this resists gaming.** The model classifies evidence; Python computes
  185. every number. Repeating a phrase twenty times collapses to one piece of
  186. evidence, a parent skill never satisfies a child requirement (Python does not
  187. imply TensorFlow), and unapproved synonyms are downgraded rather than accepted.
  188. Scores replay exactly from stored judgments with no LLM call:
  189. ```bash
  190. python evidence_match.py --replay match.json
  191. ```
  192. ### Choosing a model
  193. Because the model only classifies and Python computes every number, the model
  194. is configuration rather than policy — a weaker judge produces measurably wrong
  195. labels, never quietly drifting scores. Set either role with a
  196. `provider:model` spec; a bare name still means Anthropic.
  197. ```bash
  198. # Careful extraction, cheap judging — the split that trades cost against risk
  199. export EVIDENCE_EXTRACTOR_MODEL="anthropic:claude-sonnet-5"
  200. export EVIDENCE_JUDGE_MODEL="deepseek:deepseek-chat"
  201. # Or run entirely on self-hosted weights; nothing leaves the machine
  202. export EVIDENCE_JUDGE_MODEL="local:qwen3"
  203. export EVIDENCE_LOCAL_BASE_URL="http://localhost:11434/v1"
  204. ```
  205. Providers: `anthropic`, `xai`, `moonshot`, `deepseek`, `dashscope` (Qwen,
  206. Singapore), `dashscope-cn` (Qwen, Beijing — cheaper, but candidate data leaves
  207. the region), `openrouter`, `together`, `fireworks`, `groq`, `local` (vLLM /
  208. Ollama / llama.cpp). Each reads its own key — `DEEPSEEK_API_KEY`,
  209. `XAI_API_KEY`, `DASHSCOPE_API_KEY`, and so on. Defaults are unchanged, so an
  210. existing install keeps behaving exactly as before.
  211. A model that cannot stop reasoning is a poor fit for the judge role whatever
  212. its sticker price: judging is bounded classification, and reasoning tokens bill
  213. at the output rate, which is already ~80% of the cost here.
  214. **Using an aggregator? Pin the backend.** One OpenRouter model slug can be
  215. served by several backends at different quantizations, so two runs can record
  216. the same `judge_model` and be materially different models — the same failure as
  217. recording a constant model name, one layer up.
  218. ```bash
  219. export EVIDENCE_JUDGE_MODEL="openrouter:qwen/qwen3.8-max"
  220. export EVIDENCE_OPENROUTER_PROVIDER="deepinfra" # disables silent failover
  221. export EVIDENCE_OPENROUTER_QUANTIZATIONS="fp8" # optional, recorded too
  222. ```
  223. A pin becomes part of the recorded identity (`qwen/qwen3.8-max@deepinfra/fp8`)
  224. and therefore part of the cache key. Without one the model id is recorded as
  225. `…@unpinned`, so a stored score never claims a reproducibility it does not have.
  226. Aggregators are well suited to comparing candidates; pin before you trust a
  227. number, and prefer a direct provider key once a model is chosen.
  228. Extraction and judging are configured separately because they carry different
  229. risk. Extraction must quote the posting verbatim and classify hard gates, and
  230. runs **once** per match. Judging is bounded classification and runs **once per
  231. requirement** — roughly ten times the volume, and where the money goes.
  232. **The model is part of the result's identity.** `extractor_model` and
  233. `judge_model` are written into the audit record and hashed into the cache key,
  234. so two models never share a cache entry and a cheap-model score can never be
  235. served as a frontier-model one. Changing either invalidates cached results by
  236. design.
  237. **Before switching, measure.** Zero adversarial inversions is what makes this
  238. engine better than keyword matching, so re-run the ablations and compare
  239. against the current baseline rather than assuming a cheaper model holds:
  240. ```bash
  241. RUN_LIVE_LLM=1 python -m benchmarks.evidence.runner --with-llm
  242. ```
  243. > Routing resumes to a third-party provider sends candidate data outside
  244. > Anthropic. `pii_redactor` runs before every hosted call regardless of
  245. > provider, and `local:` keeps everything on your own machine — but the choice
  246. > of endpoint is a data-handling decision, not just a pricing one.
  247. ---
  248. ## Legacy Scoring System (diagnostic only)
  249. Retained for the comparison table and existing API integrations.
  250. ### ATS Scorer — 8 Weighted Components
  251. Heuristic keyword/semantic match scoring. **It does not simulate any real ATS.** There is no universal ATS algorithm — Workday, Greenhouse, Taleo, and other systems parse, search, filter, and rank differently, and modern products increasingly use requirement-based semantic matching rather than raw keyword counts. Treat this score as an advisory match diagnostic, never as a pass/fail prediction.
  252. | Component | Weight | What It Measures |
  253. |-----------|--------|------------------|
  254. | Phrase Match | 25% | Multi-word industry phrases present in both texts |
  255. | Keyword Match | 20% | Lemmatized keywords with synonym expansion |
  256. | Weighted Industry Terms | 15% | Domain-specific terminology with recency decay |
  257. | Semantic Similarity | 10% | SBERT vector cosine similarity between resume and JD |
  258. | BM25 Score | 10% | Probabilistic relevance ranking (BM25Plus) |
  259. | Job Title Match | 10% | Exact JD title in resume header/summary |
  260. | Graph Centrality | 5% | Infers missing skills from related skills via NetworkX |
  261. | Skill Recency | 5% | Exponential decay — recent experience weighted higher |
  262. **Additional checks:** Hidden text detection, readability analysis (Flesch-Kincaid Grade 10-12 optimal), format risk assessment.
  263. ### HR Scorer — 6 Factors + Visual Analysis
  264. Applies recruiter-inspired heuristics as an advisory review of experience and presentation signals. **It does not predict how any actual recruiter will react** — it is a structured checklist, not a behavioral model.
  265. | Factor | Weight | What It Measures |
  266. |--------|--------|------------------|
  267. | Job Fit | 15-30% | Domain/therapeutic area, experience type, education, role level |
  268. | Experience Fit | 10-20% | Years of experience vs. JD requirements, Goldilocks zone |
  269. | Skills Match | 10-30% | Demonstrated skills (action verbs) vs. listed skills |
  270. | Career Trajectory | 10-15% | Title progression via linear regression slope |
  271. | Impact Signals | 15-25% | Metrics density + Bloom's Taxonomy verb power levels |
  272. | Competitive Edge | 10% | Company/university prestige signals |
  273. | F-Pattern Visual | +/-5pts | Layout heuristics (golden triangle, left-rail alignment) |
  274. Weights adapt to detected seniority (junior / mid / senior / executive / career-pivot); ranges shown. **Risk penalties:** Job hopping (-8 to -15 pts), unexplained gaps (-5 to -15 pts), recent instability.
  275. ### LLM Scorer (Optional)
  276. Claude-powered rubric evaluation that catches nuances the algorithmic scorers miss — tone, coherence, storytelling quality.
  277. ---
  278. ## Pricing
  279. | Tier | Price | What You Get |
  280. |------|-------|-------------|
  281. | **Free** | $0 | 5 cloud scores (then automatic local scoring fallback for CLI/MCP users) |
  282. | **Pro** | $12/mo | Unlimited checks, full keyword gap + deep AI analysis, 10 AI rewrites/mo, 30 cover letters/mo |
  283. | **Ultra** | $29/mo | Everything in Pro + 100 AI rewrites/mo, 1,000 cover letters/mo |
  284. **Note for Claude Code / claude.ai users:** Your Anthropic subscription already handles resume writing via Claude. The scorer server only does ATS + HR scoring, so **Pro is all you need** — you do not need Ultra.
  285. Sign up at [getresumehq.com](https://getresumehq.com). After signing up, run `/resume-builder:setup` to link your Pro account in one step.
  286. ---
  287. ## Architecture Overview
  288. ```
  289. ┌─────────────────────────────────────────────────────────────┐
  290. │ Claude Code / claude.ai │
  291. │ /resume /tailor-resume /cover-letter /find-jobs /setup │
  292. ├─────────────────────────────────────────────────────────────┤
  293. │ │
  294. │ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌───────────┐ │
  295. │ │ ATS │ │ HR │ │ LLM │ │ Writing │ │
  296. │ │ Scorer │ │ Scorer │ │ Scorer │ │ Coach │ │
  297. │ │ (8-comp) │ │ (6-fact) │ │ (Claude) │ │ (10 rules)│ │
  298. │ └────┬─────┘ └────┬─────┘ └────┬─────┘ └─────┬─────┘ │
  299. │ └──────────────┴─────────────┘ │ │
  300. │ │ │ │
  301. │ ┌───────┴───────┐ ┌──────┴─────┐ │
  302. │ │ MCP Server │ │ DOCX │ │
  303. │ │ (FastMCP 3) │ │ Generator │ │
  304. │ │ Cloud-first │ │ (Workday) │ │
  305. │ └───────┬───────┘ └────────────┘ │
  306. │ │ │
  307. │ ┌─────────┴──────────┐ │
  308. │ │ Cloud API │ │
  309. │ │ resume-scorer │ │
  310. │ │ .fly.dev │ │
  311. │ │ (JWT + API key) │ │
  312. │ └────────────────────┘ │
  313. │ │
  314. ├─────────────────────────────────────────────────────────────┤
  315. │ Job Discovery: Adzuna + Remotive + JSearch → light score → │
  316. │ full ATS+HR score → ranked results │
  317. ├─────────────────────────────────────────────────────────────┤
  318. │ Orchestration State (state.json) — Multi-Agent DAG │
  319. │ Application Tracker (Excel) — Auto-updated per run │
  320. └─────────────────────────────────────────────────────────────┘
  321. ```
  322. The MCP server operates in **thin client** mode: it tries the cloud API first for scoring, and falls back to local scoring if the cloud is unavailable or not configured. LLM scoring always runs locally using your own API key (BYOK).
  323. ---
  324. ## Workflow
  325. ```
  326. 1. /resume-builder:setup One-time setup (install deps, create config, link Pro)
  327. 2. Create your master resume YOUR_MASTER_RESUME.md with full work history
  328. 3. /resume-builder:find-jobs [JD] Optional — discover matching jobs scored by fit
  329. 4. /resume-builder:resume [JD] Paste a job description — get a full application
  330. 5. /resume-builder:writing-coach Optional — audit and improve writing quality
  331. ```
  332. Each resume command follows a gated workflow:
  333. - **Phase 0:** Deterministic candidate-fit preflight against the configured master
  334. and exact JD (fit bar, default 50; zero hard knockouts). Rejected JDs create no output.
  335. - **Phase 1:** Read-only master/JD planning; prior tailored resumes are not inputs.
  336. - **Phase 2:** Native Researcher → Writer → Auditor → bounded Editor workflow.
  337. - **Phase 3:** Advisory ATS/HR scoring and cover-letter generation where requested.
  338. - **Phase 4:** Evidence, human-voice, and canonical-integrity authorization votes.
  339. - **Phase 5:** Ordered resume DOCX → cover-letter DOCX → tracker finalization.
  340. - **Phase 6:** Artifact verification, cleanup, and report.
  341. ---
  342. ## Alternative: pip install
  343. The scoring engine and MCP server are on PyPI:
  344. ```bash
  345. pip install resumehq
  346. # Serve the scorer to Claude/any MCP client over stdio:
  347. resumehq-mcp
  348. ```
  349. ## Alternative: Clone & Run Locally
  350. If you prefer not to use the plugin system:
  351. ```bash
  352. git clone https://github.com/jananthan30/Resume-Builder.git
  353. cd Resume-Builder
  354. pip install -r requirements.txt
  355. # Download NLTK data (one-time)
  356. python -c "import nltk; nltk.download('wordnet'); nltk.download('punkt_tab')"
  357. cp .env.example .env
  358. cp config.example.json config.json
  359. ```
  360. Then edit `.env` (API keys) and `config.json` (your info), and use commands without the `resume-builder:` prefix (e.g., `/resume` instead of `/resume-builder:resume`).
  361. ---
  362. ## Cloud Scoring API
  363. The scoring API is hosted at `https://resume-scorer.fly.dev`. Free users get **5 scored resumes**, then local scoring activates automatically. Sign up or upgrade at [getresumehq.com](https://getresumehq.com).
  364. The easiest way to link your account is via the setup wizard:
  365. ```
  366. /resume-builder:setup
  367. ```
  368. Or manually add to your `.env`:
  369. ```bash
  370. SCORER_CLOUD_URL=https://resume-scorer.fly.dev
  371. SCORER_CLOUD_API_KEY=rb_your_api_key_here
  372. ```
  373. ### MCP Configuration
  374. The `.mcp.json` file configures the MCP server to auto-start with Claude Code:
  375. ```json
  376. {
  377. "mcpServers": {
  378. "ai-resume-tuner": {
  379. "command": "python",
  380. "args": ["mcp_scorer.py"],
  381. "cwd": "/path/to/Resume-Builder",
  382. "env": {
  383. "SCORER_CLOUD_URL": "https://resume-scorer.fly.dev"
  384. }
  385. }
  386. }
  387. }
  388. ```
  389. **Environment variables:**
  390. | Variable | Required | Default | Description |
  391. |----------|----------|---------|-------------|
  392. | `SCORER_CLOUD_URL` | No | `https://resume-scorer.fly.dev` | Cloud scoring API URL |
  393. | `SCORER_CLOUD_API_KEY` | No | — | Your cloud API key (`rb_...`). Anonymous scoring (5 free) works without this. |
  394. | `ANTHROPIC_API_KEY` | No | — | For LLM scoring (always runs locally with your key) |
  395. | `ADZUNA_APP_ID` | No | — | For job discovery (free at [developer.adzuna.com](https://developer.adzuna.com/)) |
  396. | `ADZUNA_APP_KEY` | No | — | For job discovery |
  397. | `RAPIDAPI_KEY` | No | — | Enables the JSearch job source (aggregated boards) |
  398. ---
  399. ## Your Master Resume
  400. Create a file with your complete work history. Supported formats: `.docx`, `.pdf`, `.md`, or `.txt`. This is the single source of truth — all tailored resumes are generated from it. DOCX is recommended since most people already have their resume in that format.
  401. ```
  402. FULL NAME, CREDENTIALS
  403. City, State ZIP | Phone | Email | LinkedIn
  404. PROFESSIONAL SUMMARY
  405. [Your comprehensive summary with all skills and experience]
  406. PROFESSIONAL EXPERIENCE
  407. JOB TITLE | COMPANY NAME | City, State
  408. Month Year – Month Year
  409. • Achievement with quantified impact
  410. • Another achievement with metrics
  411. EDUCATION
  412. Degree Name
  413. University Name, City, State | Year – Year
  414. CERTIFICATIONS
  415. • Certification Name – Issuing Body
  416. ```
  417. Set the path to this file in your `config.json` as `master_resume_path`.
  418. ---
  419. ## Scoring Reference
  420. ### ATS Score
  421. Heuristic keyword/semantic match diagnostic. It measures textual overlap with one job description — it does **not** predict whether any employer's system will advance your resume.
  422. | Score | Rating | Meaning |
  423. |-------|--------|---------|
  424. | 80-100% | Excellent | Very high keyword/phrase overlap with this JD |
  425. | 65-79% | Good | Strong textual match with this JD |
  426. | 50-64% | Fair | Moderate match — several JD terms missing |
  427. | 35-49% | Low | Weak textual match — many JD terms missing |
  428. | 0-34% | Poor | Little keyword overlap with this JD |
  429. ### HR Score
  430. Recruiter-inspired heuristic review. The recommendation labels describe heuristic factor strength only — not any recruiter's actual decision.
  431. | Score | Recommendation | Meaning |
  432. |-------|---------------|---------|
  433. | 85%+ | STRONG INTERVIEW | Strong across heuristic factors |
  434. | 70-84% | INTERVIEW | Competitive across heuristic factors |
  435. | 55-69% | MAYBE | Marginal across heuristic factors |
  436. | <55% | PASS | Weak across heuristic factors |
  437. ---
  438. ## API Reference
  439. The scoring API runs locally (`python scorer_server.py --port 8100`) or is hosted at `https://resume-scorer.fly.dev`.
  440. ### Scoring Endpoints
  441. | Endpoint | Method | Auth | Description |
  442. |----------|--------|------|-------------|
  443. | `/health` | GET | No | Server health and version info |
  444. | `/api/match` | POST | Yes | **Evidence match — requirement-level results with exact excerpts (recommended)** |
  445. | `/score/ats` | POST | Yes | Legacy keyword/semantic match scoring (8 weighted components) |
  446. | `/score/hr` | POST | Yes | Recruiter-style heuristic scoring |
  447. | `/score/both` | POST | Yes | Legacy ATS + HR combined in one call (JSON by default, SSE with `Accept: text/event-stream`) |
  448. | `/score/llm` | POST | Yes | LLM scoring via Claude |
  449. | `/score/combined` | POST | Yes | All 3 blended (70% rules / 30% LLM) |
  450. | `/score/batch` | POST | Yes | Score multiple resume/JD pairs |
  451. | `/explain` | POST | Yes | Detailed score explanation |
  452. | `/jobs/discover` | POST | Yes | Search jobs + score against resume |
  453. ### Auth & Billing Endpoints
  454. | Endpoint | Method | Description |
  455. |----------|--------|-------------|
  456. | `/auth/register` | POST | Create account (email + password) |
  457. | `/auth/login` | POST | Login and get JWT token |
  458. | `/auth/api-key` | POST | Create an API key (requires JWT) |
  459. | `/auth/usage` | GET | Check usage stats and remaining scores |
  460. | `/billing/checkout` | POST | Start Stripe checkout for Pro upgrade |
  461. | `/billing/portal` | POST | Stripe customer portal |
  462. ### Authentication
  463. - **JWT Bearer token:** `Authorization: Bearer <token>` (from `/auth/login`)
  464. - **API key:** `X-API-Key: rb_...` (from `/auth/api-key` or web dashboard)
  465. ### Example: Score a Resume
  466. ```bash
  467. curl -X POST https://resume-scorer.fly.dev/score/ats \
  468. -H "X-API-Key: rb_your_api_key" \
  469. -H "Content-Type: application/json" \
  470. -d '{"resume_text": "Your resume text...", "jd_text": "Job description text..."}'
  471. ```
  472. ### Example: Discover Jobs
  473. ```bash
  474. curl -X POST https://resume-scorer.fly.dev/jobs/discover \
  475. -H "X-API-Key: rb_your_api_key" \
  476. -H "Content-Type: application/json" \
  477. -d '{"resume_text": "Your resume...", "job_title": "Data Scientist", "location": "New York", "max_results": 10}'
  478. ```
  479. ---
  480. ## Domain-Specific Scoring
  481. The ATS scorer auto-detects the job domain and applies domain-specific adjustments:
  482. | Domain | Detection Method | Key Adjustments |
  483. |--------|------------------|-----------------|
  484. | **Clinical Research** | SBERT prototype embeddings | Publications bonus, transferable skills mapping |
  485. | **Pharma/Biotech** | Keyword + semantic hybrid | Regulatory terminology weighting, pipeline experience |
  486. | **Technology** | Keyword + semantic hybrid | Portfolio links bonus, 1.3x skill recency weight |
  487. | **Finance** | Keyword + semantic hybrid | Deal artifacts required, 1.5x prestige weight |
  488. | **Consulting** | Keyword + semantic hybrid | Impact metrics required, 1.4x prestige weight |
  489. | **Healthcare** | Keyword + semantic hybrid | Certifications required, quality improvement focus |
  490. ---
  491. ## Supported Professions
  492. Works for **any profession**. The scorer auto-detects domain and applies appropriate weights:
  493. | Domain | Example Roles |
  494. |--------|---------------|
  495. | **Clinical Research** | CRA, Medical Monitor, Study Director, Clinical Operations |
  496. | **Pharma/Biotech** | Regulatory Affairs, Medical Science Liaison, Drug Safety |
  497. | **Technology** | Software Engineer, Product Manager, Data Scientist, ML Engineer |
  498. | **Finance** | Investment Analyst, Financial Controller, Risk Manager |
  499. | **Consulting** | Management Consultant, Strategy Analyst, Business Advisor |
  500. | **Healthcare** | Nurse Manager, Quality Director, Health Administrator |
  501. | **General** | Any role not matching a specific domain — uses universal scoring |
  502. ---
  503. ## ATS-Compliant DOCX Output
  504. The DOCX generator produces files optimized for Applicant Tracking Systems (Workday, Taleo, Greenhouse, Lever):
  505. - **No tables, text boxes, columns, or graphics** (ATS parsers can't read these)
  506. - **Heading styles** for section detection (Workday XML parsing)
  507. - **Safe fonts**: Calibri, Arial, Times New Roman (10-12pt body)
  508. - **Clean structure**: Contact info in body (not headers/footers)
  509. - **Bold metrics** for visual impact during human review
  510. ---
  511. ## Tech Stack
  512. | Component | Technology |
  513. |-----------|------------|
  514. | AI Agent Framework | [Claude Code](https://docs.anthropic.com/en/docs/claude-code) / [claude.ai](https://claude.ai) |
  515. | LLM | [Claude](https://www.anthropic.com/claude) (Anthropic) |
  516. | MCP Server | [FastMCP 3.0](https://gofastmcp.com/) (auto-starts with plugin, cloud-first thin client) |
  517. | Embeddings | [Sentence Transformers](https://sbert.net/) (all-MiniLM-L6-v2) |
  518. | NLP | NLTK (lemmatization), TextStat (readability) |
  519. | Search | BM25Plus (rank-bm25), NetworkX (skill graphs) |
  520. | Job Discovery | Adzuna API + Remotive API + JSearch (RapidAPI) |
  521. | API Server | FastAPI + Uvicorn |
  522. | Cloud Hosting | [Fly.io](https://fly.io) (auto-stop/start, persistent volume) |
  523. | Auth | JWT (PyJWT) + SQLite-backed API keys |
  524. | Billing | [Stripe](https://stripe.com) (subscription management) |
  525. | Document Generation | python-docx |
  526. | PDF Parsing | pdfplumber |
  527. | Tracking | openpyxl (Excel) |
  528. ---
  529. ## Project Structure
  530. ```
  531. Resume-Builder/
  532. ├── agents/ # Claude plugin-installed Researcher/Writer/Auditor/Editor definitions
  533. ├── .codex/agents/ # Native Codex Researcher/Writer/Auditor/Editor definitions
  534. ├── .claude/agents/ # Native Claude Code equivalents
  535. ├── .claude-plugin/ # Plugin manifest
  536. │ └── plugin.json # Plugin metadata (name, version, author)
  537. ├── .codex-plugin/ # Codex plugin manifest
  538. │ └── plugin.json # Codex metadata and install-surface copy
  539. ├── .agents/plugins/
  540. │ └── marketplace.json # Local Codex marketplace entry
  541. ├── skills/resume-team/ # Installable Codex Resume Team entrypoint
  542. ├── commands/ # Slash commands (plugin format)
  543. │ ├── setup.md # One-time setup wizard
  544. │ ├── job-fit.md # Deterministic master-vs-JD candidate-fit gate
  545. │ ├── resume.md # Full application (native four-role team)
  546. │ ├── resume-team.md # Shared fail-closed coordinator protocol
  547. │ ├── tailor-resume.md # Resume only
  548. │ ├── cover-letter.md # Cover letter only
  549. │ ├── find-jobs.md # Job discovery + scoring
  550. │ ├── batch-resume.md # Batch processing
  551. │ └── writing-coach.md # Human Voice + Impact rules (0-16)
  552. ├── hooks/ # Plugin hooks
  553. │ └── hooks.json # SessionStart: checks if scoring is ready
  554. ├── .mcp.json # MCP server config (auto-starts scorer)
  555. ├── .codex.mcp.json # Codex MCP server config
  556. ├── mcp_scorer.py # MCP scoring server (7 production-supported surfaces)
  557. ├── job_discovery.py # Job search + two-tier scoring (Adzuna + Remotive + JSearch)
  558. ├── data/ # Reference databases for scoring
  559. │ ├── keywords_*.json # Domain-specific keyword databases (6 domains)
  560. │ ├── skill_taxonomy.json # Skill categories with decay constants
  561. │ ├── company_prestige.json # Company prestige scoring
  562. │ ├── university_rankings.json# University prestige scores
  563. │ ├── acronyms.json # Industry acronym expansion
  564. │ └── action_verbs.json # Verb power classifications
  565. ├── ats_scorer.py # ATS scoring engine (2,800+ lines)
  566. ├── hr_scorer.py # HR scoring engine (2,900+ lines)
  567. ├── llm_scorer.py # LLM-powered rubric scorer
  568. ├── scorer_server.py # FastAPI REST API (v3.0 — auth, usage, billing)
  569. ├── pii_redactor.py # PII redaction via Presidio (pre-LLM API calls)
  570. ├── docx_generator.py # ATS/Workday-compliant DOCX generator
  571. ├── orchestration_state.py # Multi-agent state management (DAG)
  572. ├── multi_agent_team.py # Vendor-neutral, offline, fail-closed team controller
  573. ├── candidate_fit_preflight.py # Deterministic fit-bar/no-knockout first gate
  574. ├── native_resume_team.py # Hardened Codex/Claude CLI adapter and draft publisher
  575. ├── schemas/
  576. │ ├── resume-team-handoff.schema.json # Strict public role handoff contract
  577. │ ├── resume-team-authorization.schema.json # Three-vote authorization contract
  578. │ ├── resume-team-final-receipt.schema.json # Durable draft-authorization sidecar contract
  579. │ └── resume-team-result.schema.json # Draft-stage runtime result contract
  580. ├── tracker_utils.py # Excel application tracker utilities
  581. ├── resume_builder.py # Retired direct-rewrite CLI; native-team migration guard
  582. ├── requirements.txt # Python dependencies
  583. ├── config.example.json # Config template
  584. ├── .env.example # Environment variable template
  585. ├── AGENTS.md # Project context for Codex
  586. ├── CLAUDE.md # Project context for Claude Code
  587. ├── LICENSE # MIT License
  588. └── README.md # You are here
  589. ```
  590. ---
  591. ## Native Resume Team
  592. Codex and Claude Code use the same `resume-team/v2` control flow without API
  593. keys or a third-party orchestration framework. Project custom-agent role files
  594. omit model pins and follow their host's inheritance rules. The hardened runtime
  595. does not inherit transient parent-session or user configuration: by default its
  596. managed CLI model/reasoning selection is unknown and must not be described as a
  597. specific model, profile, or Ultra setting.
  598. - In an installed Claude Code plugin, run `/resume-builder:resume-team [JD]`;
  599. the four roles load from the plugin-root `agents/` directory.
  600. - In Codex, run `$resume-team [JD]`. The skill uses
  601. `python native_resume_team.py --host codex` as the authoritative production
  602. path; each role runs from an empty temporary working directory with tool
  603. surfaces disabled. A project checkout also registers the four read-only
  604. custom roles from `.codex/agents/` for interactive inspection, but those
  605. manual roles are not the capability-isolated publication path.
  606. - The hardened production runtime requires macOS or Linux, Python 3.10+,
  607. `config.json`, the configured master resume, `candidate_fit_preflight.py`, and
  608. the local deterministic audit helpers. Windows preflight fails closed with
  609. `POSIX_RUNTIME_REQUIRED`. No
  610. external model API key is required for the role agents.
  611. - Codex model selection is unpinned by default. Only when the user explicitly
  612. requests it may the runtime receive `--model <exact-model>` and/or
  613. `--reasoning-effort ultra`; it has no profile option. These Codex-only flags
  614. must not be passed to Claude.
  615. 0. Before any role/team invocation or output creation, the coordinator runs
  616. `candidate_fit_preflight.py` against the exact JD and only the configured
  617. master resume—never a prior tailored resume. The canonical
  618. `candidate-fit-policy-v3` report must clear the fit bar (default 50) with
  619. trustworthy extraction, zero hard knockouts, `passed: true`, and no codes. Scores below the bar
  620. (including 60–69) or hard knockouts return `REJECTED:CANDIDATE_FIT`;
  621. unavailable, malformed, stale, or mismatched reports return
  622. `FAILED:CANDIDATE_FIT_PREFLIGHT`. No automatic or manual workflow bypass exists.
  623. 1. The coordinator sends the Researcher only the job description.
  624. 2. The coordinator sends the Writer only the master resume and validated research artifact.
  625. 3. The read-only Auditor checks the exact Writer draft and cannot edit it.
  626. 4. The Editor is invoked only for named failures, with at most two corrections
  627. and a fresh audit after each edit.
  628. 5. Draft-stage publication requires the final Auditor PASS plus independent
  629. evidence, human-voice, and canonical-integrity votes on the same draft digest.
  630. Malformed, stale, replayed, ambiguous, timed-out, unavailable, side-effecting,
  631. or partially published runs fail closed. A runtime `resume-team-result/v2`
  632. `PUBLISHED` result means only
  633. that an authorized, digest-verified `resume.md` draft was atomically written and
  634. read back. It does not mean DOCX generation, tracker update, cleanup, or package
  635. completion. `/resume-builder:resume` and `/resume-builder:tailor-resume` must
  636. complete their ordered DOCX, tracker, artifact-verification, cleanup, and report
  637. gates before claiming package success; a score cannot override an authenticity
  638. gate.
  639. Every `PUBLISHED` result includes the independently reproducible
  640. `candidate_fit_report` and `candidate_fit_report_digest`, plus an inline
  641. `resume-team-final-receipt/v2` `authorization_receipt`, its canonical
  642. `authorization_receipt_digest`, and a durable `authorization_receipt_path`.
  643. The sidecar conforms to `schemas/resume-team-final-receipt.schema.json`.
  644. Downstream finalization resolves the path against the output directory when
  645. relative, requires its resolved parent to be that directory, reads only a regular
  646. non-symlink JSON sidecar, and matches its canonical digest, run/case IDs, exact
  647. passing candidate-fit report/digest, and draft and verified-target digests against
  648. the result, configured master, exact JD, and independently hashed `resume.md`.
  649. It also recomputes the master `source_digest` from `config.json`, recomputes
  650. `job_description_digest` from the fixed sibling `job_description.txt`, and requires
  651. a SHA-256 Researcher artifact plus distinct same-host native Researcher/Auditor IDs.
  652. The receipt must also carry a same-draft PASS `auditor_attestation` and the complete
  653. passing `authorization_report`: no codes, exactly three ordered named PASS votes on
  654. the same draft with distinct IDs, `canonical_digest(report) ==
  655. authorization_digest`, and an identical ordered `vote_invocation_ids` list. The
  656. same check is repeated immediately before DOCX generation. Cleanup preserves the
  657. receipt as durable audit evidence.
  658. Finalization is code-bound: callers retain the captured runtime result, invoke
  659. `final_receipt_verifier.py` with its exact receipt path, digest, and config, and use only
  660. `create_resume_from_md_authorized`, `create_cover_letter_from_md_authorized`, and
  661. `add_application_authorized`. Each wrapper revalidates authorization at the
  662. side-effect boundary; tracker success requires a literal `True` return.
  663. The constructive-provenance experiment established that a self-consistent
  664. model-supplied evidence ledger is not a trust root. Such a ledger is accepted only
  665. when its digest is independently attested. Production therefore anchors every
  666. changed line directly to coordinator-attested, same-role master-resume spans and
  667. applies the closed lexical verifier. `constructive_provenance.py` is a conditional
  668. checker and test artifact, not an alternative publication path.
  669. Claude role definitions and the native runtime use an explicit zero-tool allowlist, so they cannot
  670. actively inspect workspace files; Claude Code may still supply its normal
  671. project startup instructions and basic environment context. Codex custom agents
  672. use a read-only sandbox, which prevents writes but is not a filesystem-read
  673. isolation boundary. In both cases, scoped payloads describe coordinator data
  674. flow rather than every byte of host-provided context. Codex currently has no
  675. documented per-custom-agent built-in-tool denylist, so manual Codex role instructions
  676. prohibit unrelated reads and must not be represented as capability isolation.
  677. ## Writing Coach — Rules 0-16
  678. The `/writing-coach` command applies human-voice and impact rules to every bullet point. Core rules include:
  679. 1. **Plain Verb Start** — Use direct verbs such as Led, Built, Wrote, Cut, Reviewed, or Directed; AI-cliché openers are banned
  680. 2. **Quantified Impact** — 40%+ of bullets must contain metrics (%, $, numbers)
  681. 3. **So-What Test** — Every bullet answers "why does this matter?"
  682. 4. **Jargon Calibration** — Match terminology level to the target role
  683. 5. **Tense Consistency** — Past tense for past roles, present for current
  684. 6. **Parallel Structure** — Consistent grammatical patterns within sections
  685. 7. **Length Optimization** — 1-2 lines per bullet, no walls of text
  686. 8. **Keyword Integration** — Natural placement, never forced
  687. 9. **Achievement vs. Duty** — Frame responsibilities as accomplishments
  688. 10. **Readability** — Flesch-Kincaid Grade 10-12 target
  689. ---
  690. ## Contributing
  691. Contributions are welcome! Some ideas:
  692. - **New domain profiles** — add keyword databases for law, marketing, academia, etc.
  693. - **Additional job boards** — integrate Indeed, LinkedIn, or regional boards
  694. - **Additional ATS parsers** — test against more ATS systems (Taleo, iCIMS, Greenhouse)
  695. - **Resume templates** — add more DOCX template styles
  696. - **Internationalization** — support for non-English resumes and job markets
  697. ```bash
  698. git checkout -b feature/your-feature
  699. # ... make changes ...
  700. git commit -m "Add your feature"
  701. git push origin feature/your-feature
  702. ```
  703. ---
  704. ## License
  705. MIT License — see the [LICENSE](LICENSE) file for details.
  706. ---
  707. ## Acknowledgments
  708. - Built with [Claude Code](https://docs.anthropic.com/en/docs/claude-code) by [Anthropic](https://www.anthropic.com/)
  709. - Scoring heuristics informed by public research on how recruiting systems parse, search, and rank resumes (see `docs/`)
  710. - Domain keyword databases curated from thousands of real job descriptions
  711. - Job search powered by [Adzuna](https://www.adzuna.com/) and [Remotive](https://remotive.com/)
  712. ---
  713. **If this project helps you land interviews, give it a star ⭐**

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.