dennisthiessen/claude-resume-kit
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs: restructure README, add example output PDFs and DOCS.md

Browse Source 
- README.md: rewritten as landing page (~160 lines) — problem/solution
  hook, "what makes this different", example output links, "try it first"
  fast path, streamlined quickstart
- DOCS.md: new reference doc with architecture, concepts, customization
  tables, design decisions, and FAQ (moved from README)
- Example output: compiled resume + cover letter PDFs for Dr. Jordan Chen
  (fictional) applying to Whitfield University faculty position
- Example .tex sources in resume_builder/examples/output/
- GitHub repo description and topics updated
This commit is contained in:
Akhil Reddy Peeketi
2026-03-09 19:37:54 -06:00
parent 54f7b91aa6
commit 80fb7d6d5a
6 changed files with 494 additions and 331 deletions
Download Patch File Download Diff File Expand all files Collapse all files
DOCS.md
+211
View File
@@ -0,0 +1,211 @@
# Documentation
Detailed reference for claude-resume-kit. For the quick overview, see [README.md](README.md).
---
## Architecture
```
claude-resume-kit/
├── CLAUDE.md # Auto-loaded project instructions
├── config.md # Your personal configuration
├── .claude/skills/ # 6 slash commands
│ ├── setup-extract/SKILL.md # Extract from papers → structured data
│ ├── setup-build-kb/SKILL.md # Synthesize KB from extractions
│ ├── make-resume/SKILL.md # JD → tailored resume/CV (.tex)
│ ├── make-cl/SKILL.md # Session → cover letter (.tex)
│ ├── edit-resume/SKILL.md # Edit from critique/feedback
│ └── critique/SKILL.md # Independent quality review
├── resume_builder/
│ ├── reference/ # Generation rules and protocols
│ │ ├── shared_ops.md # Session workflow (all skills read this)
│ │ ├── resume_reference.md # Resume/CV formatting rules
│ │ ├── cl_reference.md # Cover letter rules
│ │ ├── critical_rules.md # Compact re-read for generation phase
│ │ ├── session_file_template.md # Session file format spec
│ │ └── critique_framework.md # 8-part critique system
│ ├── templates/ # LaTeX .cls classes + .tex templates
│ │ ├── resume.cls # 2-page resume class
│ │ ├── cv.cls # Multi-page CV class
│ │ ├── resume_template.tex # Resume structural template
│ │ ├── cv_template.tex # CV structural template
│ │ └── coverletter_template.tex # Cover letter template
│ ├── helpers/
│ │ └── char_count.py # Character counting utility for bullets
│ ├── examples/ # Fictional "Dr. Jordan Chen" — full worked example
│ ├── experience/ # YOUR experience files (built by /setup-build-kb)
│ ├── bundles/ # YOUR role-type bundles (built by /setup-build-kb)
│ └── support/ # Skills taxonomy, pub metadata, AI fingerprint rules
├── knowledge_base/
│ ├── extractions/ # Paper extractions (built by /setup-extract)
│ ├── papers/ # Drop your PDFs / .tex source here
│ └── notes/ # Any other reference material
├── JDs/ # Job descriptions (text files)
└── output/ # Generated .tex files, session files, critiques
```
---
## Concepts
### Session Files
Every JD gets a session file (`output/<Folder>/session_<name>.md`) that tracks:
- JD analysis and ATS keywords
- Which bundle was selected
- Bullet plan (which achievements, in what order, at what length)
- All generation decisions and their rationale
- Cover letter plan
- Critique scores
All 4 generation skills read and update this file. It's the single source of truth for each application.
### Experience Files
One file per position (e.g., `experience_postdoc_university.md`). Each achievement has:
- **Source paper** with citation
- **Methods and tools** used
- **Quantitative results**
- **Pre-written bullet variants** (2-line and 3-line)
- **Tags** for which role types this achievement is relevant to
- **Significance** context for cover letters
### Role-Type Bundles
One file per target audience (e.g., `bundle_academic.md`). Each bundle contains:
- **S1: Role Profile** — what this audience values, positioning strategy
- **S2: Summary Guide** — how to write the summary for this role type
- **S3: Achievement Reframing Map** — priority ranking of your achievements for this audience
- **S4: Skills Guide** — which tools to bold, which to include, grouping strategy
- **S5: Cover Letter Guide** — opening hooks, paragraph templates, anti-patterns
### Provenance Flags
The system enforces accuracy through provenance tracking in `config.md`. Every achievement is tagged with its publication status. The skills check this table before every output and will never:
- Claim unpublished work is published
- Claim internal tools are peer-reviewed
- Use full-ownership verbs for shared work
- Inflate author position
### The Critique System
The `/critique` skill runs a multi-part assessment:
1. **Domain-Specialist Lens** — reviewer persona, gap analysis, competitive landscape
2. **Five-Perspective Read-Through** — ATS bot, recruiter (10s), HR (30s), hiring manager (2min), technical reviewer (10min)
3. **Eight-Dimension Scoring** — weighted score out of 100
4. **Interview Likelihood** — per-reader probability estimates
5. **Tiered Improvements** — ranked by point impact
6. **Interview Bridge Points** — resume-to-interview talking points
7. **Cover Letter Critique** — 6 sub-checks (anti-patterns, tailoring, context-specific, ATS keywords, structural, package cohesion)
8. **Post-Generation Verification** — mechanical and content checklists including AI fingerprint scan
---
## Three-Session Workflow
For best results, use a **separate Claude Code session** for each step. This gives each skill fresh context, which produces better quality (especially for critique — you want fresh eyes, not the same context that generated the resume).
```
Session 1: /make-resume JDs/job.txt → resume/CV .tex
/clear
Session 2: /make-cl → cover letter .tex
/clear
Session 3: /critique → critique .md with score
/clear
/edit-resume → refined .tex (if needed)
```
---
## Customization
### Everything in `config.md` (edit directly)
| Setting | What it controls | Example |
|---------|-----------------|---------|
| **Personal Info** | Name, email, phone, links on all outputs | Your contact details |
| **Document Preferences** | Page counts, bullet line variants, skills layout | `Resume: 2 pages, CV: 5 pages` |
| **Provenance Flags** | What claims are safe to make | `ML paper: under review → never say "published"` |
| **Role Types** | Target audiences and their bundles | `Academic (Tier 1), Industry R&D (Tier 2)` |
| **Decision Tree** | How JD keywords map to role types | `"tenure-track" → Academic` |
| **FIXED Sections** | Template sections that never change per JD | `Education, Publications, Awards` |
| **Output Rules** | Package formats and constraints | `Resume: 2pg + 1pg CL = 3pg package` |
| **KB Corrections** | Errors to never re-introduce | `Spearman is 0.82, not 0.85` |
### LaTeX Templates (edit directly)
- **Fonts, colors, spacing** — modify `.cls` files
- **Section order** — reorder sections in `.tex` templates
- **FIXED content** — fill in education, awards, publications, header
- **Icons** — replace `GS.png` / `orcid.png` with your own
- **Page geometry** — adjust margins in `.cls` if needed
### Knowledge Base (built by skills, then editable)
| File | How to customize |
|------|-----------------|
| **Experience files** | Edit bullet text, add/remove achievements, adjust tags |
| **Bundles** | Change priority matrices, rewrite summary guides, add role types |
| **Skills taxonomy** | Add/remove skills, change groupings, adjust bold rules |
| **Pub metadata** | Update citation counts, add new publications |
### Reference Docs (advanced)
| File | What you'd change |
|------|-------------------|
| `resume_reference.md` | Page budgets, character limits, section specs |
| `cl_reference.md` | Cover letter paragraph templates, word count targets |
| `critical_rules.md` | Generation-time rules tables |
| `critique_framework.md` | Scoring weights, critique dimensions |
| `shared_ops.md` | Session workflow, file derivation logic |
### Skill Prompts (advanced)
Each skill is a markdown file in `.claude/skills/<name>/SKILL.md`. You can:
- Add STOP points for more user control
- Change the number of web searches in Phase 0
- Adjust how many bullets per position
- Modify the critique scoring weights
- Add new skills for your workflow
---
## Key Design Decisions
- **Accuracy > Relevance > Impact > ATS > Brevity** — the priority hierarchy for every generation decision
- **LaTeX-only output** — Claude generates `.tex`, you compile locally. No formatting surprises.
- **FLIPPED position format** — the bold line under each position title is a JD-customized theme, not a generic description. This is the strongest tailoring lever.
- **Structured provenance** — every achievement is tracked from source paper through extraction to experience file to resume bullet
- **Character-precise budgets** — every bullet is calibrated to fit the template geometry, not "try to keep it short"
- **Session files as state** — all decisions for a JD live in one file. Skills can recover from interruptions.
- **Anti-fabrication by design** — provenance flags, verb discipline, and corrections logs prevent overclaiming even under pressure to impress
- **AI fingerprint avoidance** — a dedicated rules file is loaded by all generation and critique skills, covering banned words and phrases (with technical exceptions), structural anti-patterns, positive markers, and a 12-item post-generation checklist
---
## FAQ
**Q: Do I need to know LaTeX?**
No. Claude generates the `.tex` files. You just compile them (`pdflatex file.tex`). The templates handle all formatting.
**Q: How many papers should I extract?**
All papers where you're first author or co-first author, plus key contributing-author papers. Quality matters more than quantity — 5 well-extracted papers beat 20 shallow ones.
**Q: Can I use this for non-academic roles?**
Yes. The framework supports any role type — define them in `config.md`. Industry R&D, consulting, data science, and engineering roles all work. Just create appropriate bundles.
**Q: What if I don't have a Google Scholar / ORCID?**
Remove those lines from the templates. The framework adapts to what you have.
**Q: How do I update after publishing new papers?**
Run `/setup-extract` on the new paper, then update your experience file and bundles. Existing session files are not affected.
**Q: Can I use this with resume formats other than the included templates?**
Yes. The `.cls` files define the visual style. You can modify them or write your own. The skills generate content based on the template structure — update the `[GENERATE: ...]` and `[FIXED: ...]` markers in your template.
**Q: Can multiple people use the same kit?**
Each person needs their own clone with their own `config.md`, knowledge base, and templates. The framework itself is shared; the content is personal.
**Q: What Claude model should I use?**
The skills are designed for Claude's most capable models (Opus, Sonnet). Less capable models may skip steps or produce lower-quality output.
README.md
+86 -331
View File
@@ -1,404 +1,159 @@
# claude-resume-kit # claude-resume-kit
A Claude Code-powered framework for generating tailored resumes, CVs, and cover letters from structured knowledge bases. Most AI resume tools work the same way: paste resume + paste JD, get a rewrite. They don't know which of your papers is published vs. under review. They don't know you only ran the simulations, not the experiments. They'll upgrade "contributed to" into "developed" without blinking.
Built for researchers, engineers, and technical professionals who apply to many positions and need each application package customized — without starting from scratch every time. This is different. You extract your papers, codebases, and reports once — the system asks structured questions about each one. After that, every new application is just pointing it at a JD. It picks the right achievements, frames them for the audience, enforces accuracy, and generates LaTeX you compile locally.
## Why This Exists Built for researchers and engineers with lots of source material (papers, code, reports) who apply to many positions across different employer types.
Applying to jobs as a researcher is painful. Each position needs a different framing of the same work: a national lab cares about method development, industry wants throughput metrics, academia wants publications and teaching. You end up with 15 slightly different Word docs, inconsistent formatting, and no systematic way to ensure you haven't overclaimed something.
This framework treats resume generation as a **structured pipeline** with enforced accuracy rules, not a "rewrite my resume" chat prompt. You build your knowledge base once, then generate tailored output for each JD in minutes.
**What makes it different from ChatGPT resume prompts:**
- **Anti-fabrication rules** baked into every skill — accuracy always beats impressiveness
- **Provenance tracking** — knows what's published vs. under review vs. internal
- **Role-type bundles** that frame the same work differently for different audiences
- **AI fingerprint avoidance** — banned words, structural anti-patterns, and a 12-item post-gen checklist so output reads human-written
- **Mechanical enforcement** of page budgets, character limits, and formatting rules
- **Session files** that track every decision, making edits and critiques stateful
- **LaTeX output** — pixel-perfect formatting, not "close enough"
--- ---
## How It Works ## What makes this different
``` **Knowledge base, not a rewriter.** You extract once. Every application draws from verified source material — not a pasted resume that gets "improved."
Your Papers ──→ /setup-extract ──→ Extractions ──→ /setup-build-kb ──→ Knowledge Base
│
Job Description ──→ /make-resume ──→ Tailored Resume/CV (.tex) │
│ ↓ │
/make-cl ──→ Cover Letter (.tex) │
│ ↓ │
/critique ──→ 9-Part Score + AI Scan + Fixes │
│ ↓ │
/edit-resume ──→ Refined Package │
```
**One-time setup (do once):** **Anti-fabrication by design.** Provenance flags on every achievement (published / under review / internal). Verb discipline rules prevent overclaiming. A corrections log ensures fixed errors don't reappear.
1. `/setup-extract` — Run on each paper/project to create structured extraction files
2. `/setup-build-kb` — Synthesize extractions into experience files, bundles, and skills taxonomy
**Per-application (do for each JD):** **AI fingerprint avoidance.** Banned-word lists, structural anti-patterns, and a 12-item post-generation scan so output reads as human-written.
3. `/make-resume <JD>` — Analyze JD, plan bullets, generate tailored LaTeX
4. `/make-cl` — Generate a matching cover letter from the session file **Multi-perspective critique.** Five reader personas (ATS bot through technical reviewer) score your resume across 8 dimensions in a fresh context window.
5. `/critique` — Independent 8-dimension quality review
6. `/edit-resume` — Refine based on critique or your own feedback **LaTeX output, locally compiled.** No data leaves your machine beyond the Claude Code conversation.
---
## Example Output
Here's what the system generates for the included fictional researcher (Dr. Jordan Chen, computational biologist) applying to a tenure-track faculty position:
- [Example Resume (PDF)](resume_builder/examples/example_resume.pdf) — 2-page resume with JD-tailored bullets, skills, and publications
- [Example Cover Letter (PDF)](resume_builder/examples/example_cover_letter.pdf) — 1-page academic cover letter with specific hooks
- [Example Session File](resume_builder/examples/example_session_file.md) — the decision log that produced this output
- [Source .tex files](resume_builder/examples/output/) — the LaTeX source Claude generated
All example data is in `resume_builder/examples/` — extraction, experience file, bundle, config, and session file.
---
## What you actually do
**One-time setup (~10 min per paper):**
1. Drop your papers/reports into `knowledge_base/papers/`
2. Run `/setup-extract` on each — Claude reads it and asks you questions about your contributions and publication status
3. Run `/setup-build-kb` — synthesizes everything into your knowledge base
**Per application (~15-20 min):**
1. Drop the JD into `JDs/`
2. Run `/make-resume JDs/target_job.txt` — approve the bullet plan, get a `.tex` file
3. Run `/make-cl` for a cover letter
4. Run `/critique` for a scored review with specific fixes
Each step uses a **separate Claude Code session** for best quality (fresh context = less bias). Each step uses a **separate Claude Code session** for best quality (fresh context = less bias).
--- ---
## Architecture
```
claude-resume-kit/
├── CLAUDE.md # Auto-loaded project instructions
├── config.md # Your personal configuration
├── .claude/skills/ # 6 slash commands
│ ├── setup-extract/SKILL.md # Extract from papers → structured data
│ ├── setup-build-kb/SKILL.md # Synthesize KB from extractions
│ ├── make-resume/SKILL.md # JD → tailored resume/CV (.tex)
│ ├── make-cl/SKILL.md # Session → cover letter (.tex)
│ ├── edit-resume/SKILL.md # Edit from critique/feedback
│ └── critique/SKILL.md # Independent quality review
├── resume_builder/
│ ├── reference/ # Generation rules and protocols
│ │ ├── shared_ops.md # Session workflow (all skills read this)
│ │ ├── resume_reference.md # Resume/CV formatting rules
│ │ ├── cl_reference.md # Cover letter rules
│ │ ├── critical_rules.md # Compact re-read for generation phase
│ │ ├── session_file_template.md # Session file format spec
│ │ └── critique_framework.md # 8-part critique system
│ ├── templates/ # LaTeX .cls classes + .tex templates
│ │ ├── resume.cls # 2-page resume class
│ │ ├── cv.cls # Multi-page CV class
│ │ ├── resume_template.tex # Resume structural template
│ │ ├── cv_template.tex # CV structural template
│ │ └── coverletter_template.tex # Cover letter template
│ ├── helpers/
│ │ └── char_count.py # Character counting utility for bullets
│ ├── examples/ # Fictional "Dr. Jordan Chen" — full worked example
│ ├── experience/ # YOUR experience files (built by /setup-build-kb)
│ ├── bundles/ # YOUR role-type bundles (built by /setup-build-kb)
│ └── support/ # Skills taxonomy, pub metadata, AI fingerprint rules
├── knowledge_base/
│ ├── extractions/ # Paper extractions (built by /setup-extract)
│ ├── papers/ # Drop your PDFs / .tex source here
│ └── notes/ # Any other reference material
├── JDs/ # Job descriptions (text files)
└── output/ # Generated .tex files, session files, critiques
```
---
## Prerequisites ## Prerequisites
- **[Claude Code](https://docs.anthropic.com/en/docs/claude-code)** CLI installed and authenticated - **[Claude Code](https://docs.anthropic.com/en/docs/claude-code)** CLI installed and authenticated
- **A LaTeX distribution** for compiling `.tex` → `.pdf` (e.g., [TeX Live](https://tug.org/texlive/), [MacTeX](https://tug.org/mactex/), [MiKTeX](https://miktex.org/)) - **A LaTeX distribution** for compiling `.tex` to `.pdf` (e.g., [TeX Live](https://tug.org/texlive/), [MacTeX](https://tug.org/mactex/), [MiKTeX](https://miktex.org/))
- **Your research papers** or project documentation ready for extraction - **Your research papers** or project documentation ready for extraction
--- ---
## Quickstart ## Try it first (5 minutes)
### 1. Clone and enter the project Want to see what it does before extracting your own papers? The repo includes a complete example knowledge base for a fictional researcher:
```bash
git clone https://github.com/ARPeeketi/claude-resume-kit.git
cd claude-resume-kit
claude
/make-resume JDs/example_jd.txt
```
This runs the full pipeline — JD analysis, bullet selection, LaTeX generation — using the included example data. No setup required.
---
## Full Setup
### 1. Clone and configure
```bash ```bash
git clone https://github.com/ARPeeketi/claude-resume-kit.git git clone https://github.com/ARPeeketi/claude-resume-kit.git
cd claude-resume-kit cd claude-resume-kit
``` ```
### 2. Configure your profile Edit `config.md` with your details (name, email, provenance flags, role types). See `resume_builder/examples/example_config.md` for a complete example.
Edit `config.md` with your details. This file is read by every skill. ### 2. Extract your papers
```markdown Place PDFs or `.tex` source files in `knowledge_base/papers/`, then:
## Personal Info
- **Name:** Your Name
- **Email:** you@email.com
- **LinkedIn:** linkedin.com/in/yourprofile
...
## Provenance Flags
| Item | Status | Correct Framing |
|------|--------|----------------|
| My ML paper | published in Nature | OK to say "published" |
| Internal tool | unpublished | "infrastructure I developed" — never imply peer-reviewed |
## Role Types
| Role Name | Target Employers | Tier | Bundle File |
|-----------|-----------------|------|-------------|
| Academic | R1 universities | 1 | bundle_academic.md |
| Industry | Tech companies | 2 | bundle_industry.md |
```
See `resume_builder/examples/example_config.md` for a complete example.
### 3. Drop your papers
Place PDFs or `.tex` source files in `knowledge_base/papers/`.
### 4. Extract from each paper
``` ```
/setup-extract knowledge_base/papers/my_paper.pdf /setup-extract knowledge_base/papers/my_paper.pdf
``` ```
Claude reads the paper, asks clarifying questions about your specific contributions, then creates a structured extraction in `knowledge_base/extractions/`. Repeat for each paper. Claude reads the paper, asks clarifying questions about your contributions, and creates a structured extraction. Repeat for each paper.
### 5. Build your knowledge base ### 3. Build your knowledge base
``` ```
/setup-build-kb /setup-build-kb
``` ```
This synthesizes all extractions into: This synthesizes all extractions into experience files, role-type bundles, and support files.
- **Experience files** (`resume_builder/experience/`) — one per position, with pre-written 2L and 3L bullet variants for every achievement
- **Role-type bundles** (`resume_builder/bundles/`) — positioning guides for each audience (what to emphasize, what to omit, how to frame)
- **Support files** (`resume_builder/support/`) — skills taxonomy, publication metadata, AI fingerprint avoidance rules
### 6. Customize your LaTeX templates ### 4. Customize your LaTeX templates
Open the templates and fill in your **FIXED sections** — content that never changes per JD: Open the templates in `resume_builder/templates/` and fill in your FIXED sections — education, header, awards, publications. The `[CONFIG: ...]` placeholders show you what to fill in.
- `resume_builder/templates/resume_template.tex` — Education, header, awards ### 5. Generate for a job
- `resume_builder/templates/cv_template.tex` — Education, publications, header, awards, collaborations
- `resume_builder/templates/coverletter_template.tex` — Header, signature block
The `[CONFIG: ...]` placeholders show you exactly what to fill in. The `[GENERATE: ...]` sections are filled by Claude during generation.
### 7. Generate for a job
Save the job description as a text file in `JDs/`, then:
``` ```
/make-resume JDs/target_job.txt /make-resume JDs/target_job.txt
``` ```
This runs a 3-phase pipeline: Then in separate sessions: `/make-cl` for the cover letter, `/critique` for a scored review.
- **Phase 0:** Web-searches the company, analyzes JD keywords, selects role-type bundle
- **Phase 1:** Plans which bullets to include and in what order (you approve the plan)
- **Phase 2:** Generates the full `.tex` file with enforced character limits
### 8. Cover letter + critique (separate sessions)
```
/clear
/make-cl
```
```
/clear
/critique
```
Then `/edit-resume` to address any critique findings.
--- ---
## Concepts ## How It Works
### Session Files ```
Your Papers --> /setup-extract --> Extractions --> /setup-build-kb --> Knowledge Base
Every JD gets a session file (`output/<Folder>/session_<name>.md`) that tracks: |
- JD analysis and ATS keywords Job Description --> /make-resume --> Tailored Resume/CV (.tex) |
- Which bundle was selected | v |
- Bullet plan (which achievements, in what order, at what length) /make-cl --> Cover Letter (.tex) |
- All generation decisions and their rationale | v |
- Cover letter plan /critique --> 8-Part Score + AI Scan + Fixes |
- Critique scores | v |
/edit-resume --> Refined Package |
All 4 generation skills read and update this file. It's the single source of truth for each application. ```
### Experience Files
One file per position (e.g., `experience_postdoc_university.md`). Each achievement has:
- **Source paper** with citation
- **Methods and tools** used
- **Quantitative results**
- **Pre-written bullet variants** (2-line and 3-line)
- **Tags** for which role types this achievement is relevant to
- **Significance** context for cover letters
### Role-Type Bundles
One file per target audience (e.g., `bundle_academic.md`). Each bundle contains:
- **S1: Role Profile** — what this audience values, positioning strategy
- **S2: Summary Guide** — how to write the summary for this role type
- **S3: Achievement Reframing Map** — priority ranking of your achievements for this audience
- **S4: Skills Guide** — which tools to bold, which to include, grouping strategy
- **S5: Cover Letter Guide** — opening hooks, paragraph templates, anti-patterns
### Provenance Flags
The system enforces accuracy through provenance tracking in `config.md`. Every achievement is tagged with its publication status. The skills check this table before every output and will never:
- Claim unpublished work is published
- Claim internal tools are peer-reviewed
- Use full-ownership verbs for shared work
- Inflate author position
### The Critique System
The `/critique` skill runs a 9-part assessment:
1. **Domain-Specialist Lens** — reviewer persona, gap analysis, competitive landscape
2. **Five-Perspective Read-Through** — ATS bot, recruiter (10s), HR (30s), hiring manager (2min), technical reviewer (10min)
3. **Eight-Dimension Scoring** — weighted score out of 100
4. **Interview Likelihood** — per-reader probability estimates
5. **Tiered Improvements** — ranked by point impact
6. **Interview Bridge Points** — resume-to-interview talking points
7. **AI Fingerprint Scan** — banned words, em-dash count, `-ing` endings, 12-item checklist
8. **Cover Letter Critique** — 6 sub-checks
9. **Post-Generation Verification** — mechanical and content checklists
---
## What You Can Customize
### Everything in `config.md` (edit directly)
| Setting | What it controls | Example |
|---------|-----------------|---------|
| **Personal Info** | Name, email, phone, links on all outputs | Your contact details |
| **Document Preferences** | Page counts, bullet line variants, skills layout | `Resume: 2 pages, CV: 5 pages` |
| **Provenance Flags** | What claims are safe to make | `ML paper: under review → never say "published"` |
| **Role Types** | Target audiences and their bundles | `Academic (Tier 1), Industry R&D (Tier 2)` |
| **Decision Tree** | How JD keywords map to role types | `"tenure-track" → Academic` |
| **FIXED Sections** | Template sections that never change per JD | `Education, Publications, Awards` |
| **Output Rules** | Package formats and constraints | `Resume: 2pg + 1pg CL = 3pg package` |
| **KB Corrections** | Errors to never re-introduce | `Accuracy is 2.1 meV/atom, not 2.3` |
### LaTeX Templates (edit directly)
- **Fonts, colors, spacing** — modify `.cls` files
- **Section order** — reorder sections in `.tex` templates
- **FIXED content** — fill in education, awards, publications, header
- **Icons** — replace `GS.png` / `orcid.png` with your own
- **Page geometry** — adjust margins in `.cls` if needed
### Knowledge Base (built by skills, then editable)
| File | How to customize |
|------|-----------------|
| **Experience files** | Edit bullet text, add/remove achievements, adjust tags |
| **Bundles** | Change priority matrices, rewrite summary guides, add role types |
| **Skills taxonomy** | Add/remove skills, change groupings, adjust bold rules |
| **Pub metadata** | Update citation counts, add new publications |
### Reference Docs (advanced — edit if you know what you're doing)
| File | What you'd change |
|------|-------------------|
| `resume_reference.md` | Page budgets, character limits, section specs |
| `cl_reference.md` | Cover letter paragraph templates, word count targets |
| `critical_rules.md` | Generation-time rules tables |
| `critique_framework.md` | Scoring weights, critique dimensions |
| `shared_ops.md` | Session workflow, file derivation logic |
### Skill Prompts (advanced)
Each skill is a markdown file in `.claude/skills/<name>/SKILL.md`. You can:
- Add STOP points for more user control
- Change the number of web searches in Phase 0
- Adjust how many bullets per position
- Modify the critique scoring weights
- Add new skills for your workflow
---
## Skill Reference
| Skill | Purpose | Input | Output | | Skill | Purpose | Input | Output |
|-------|---------|-------|--------| |-------|---------|-------|--------|
| `/setup-extract` | Extract structured data from a paper | Paper path | `knowledge_base/extractions/*.md` | | `/setup-extract` | Extract structured data from a paper | Paper path | `knowledge_base/extractions/*.md` |
| `/setup-build-kb` | Build resume artifacts from extractions | All extractions | `resume_builder/{experience,bundles,support}/` | | `/setup-build-kb` | Build KB from extractions | All extractions | `resume_builder/{experience,bundles,support}/` |
| `/make-resume` | Generate tailored resume or CV | JD path | `output/<Folder>/e2e_*.tex` + session file | | `/make-resume` | Generate tailored resume or CV | JD path | `output/<Folder>/e2e_*.tex` + session file |
| `/make-cl` | Generate matching cover letter | Session file | `output/<Folder>/*_cover_letter.tex` | | `/make-cl` | Generate matching cover letter | Session file | `output/<Folder>/*_cover_letter.tex` |
| `/edit-resume` | Edit resume/CV/CL from feedback | Session file + feedback | Updated `.tex` files | | `/edit-resume` | Edit resume/CV/CL from feedback | Session + feedback | Updated `.tex` files |
| `/critique` | Independent quality review | Session file | `output/<Folder>/critique_*.md` | | `/critique` | Independent quality review | Session file | `output/<Folder>/critique_*.md` |
--- ---
## Three-Session Workflow ## Documentation
For best results, use a **separate Claude Code session** for each step. This gives each skill fresh context, which produces better quality (especially for critique — you want fresh eyes, not the same context that generated the resume). For architecture details, customization tables, the full critique system breakdown, key design decisions, and FAQ, see **[DOCS.md](DOCS.md)**.
```
Session 1: /make-resume JDs/job.txt → resume/CV .tex
/clear
Session 2: /make-cl → cover letter .tex
/clear
Session 3: /critique → critique .md with score
/clear
/edit-resume → refined .tex (if needed)
```
---
## Key Design Decisions
- **Accuracy > Relevance > Impact > ATS > Brevity** — the priority hierarchy for every generation decision
- **LaTeX-only output** — Claude generates `.tex`, you compile locally. No formatting surprises.
- **FLIPPED position format** — the bold line under each position title is a JD-customized theme, not a generic description. This is the strongest tailoring lever.
- **Structured provenance** — every achievement is tracked from source paper → extraction → experience file → resume bullet
- **Character-precise budgets** — every bullet is calibrated to fit the template geometry. No "try to keep it short."
- **Session files as state** — all decisions for a JD live in one file. Skills can recover from interruptions.
- **Anti-fabrication by design** — provenance flags, verb discipline, and corrections logs prevent overclaiming even under pressure to impress.
- **AI fingerprint avoidance** — a dedicated rules file (`resume_builder/support/ai_fingerprint_rules.md`) is loaded by all 4 generation/critique skills. It includes banned words and phrases (with technical exceptions), structural anti-patterns (e.g., excessive `-ing` bullet endings, prose em-dashes), positive markers to prefer, and a 12-item post-generation checklist. Templates use period separators instead of em-dashes in Fellowships/Honors sections to avoid a common AI tell.
---
## Examples
The `resume_builder/examples/` directory contains a complete worked example for a fictional researcher, **Dr. Jordan Chen** (computational biologist). This includes:
- `example_config.md` — filled-in configuration
- `extractions/example_extraction.md` — a paper extraction
- `experience/example_experience.md` — experience file with 11 achievements across 2 positions
- `bundles/example_bundle.md` — an Academic role-type bundle
- `example_session_file.md` — a completed session file showing the full pipeline
Study these to understand the data model before building your own knowledge base.
---
## FAQ
**Q: Do I need to know LaTeX?**
No. Claude generates the `.tex` files. You just compile them (`pdflatex file.tex`). The templates handle all formatting.
**Q: How many papers should I extract?**
All papers where you're first author or co-first author, plus key contributing-author papers. Quality matters more than quantity — 5 well-extracted papers beat 20 shallow ones.
**Q: Can I use this for non-academic roles?**
Yes. The framework supports any role type — define them in `config.md`. Industry R&D, consulting, data science, and engineering roles all work. Just create appropriate bundles.
**Q: What if I don't have a Google Scholar / ORCID?**
Remove those lines from the templates. The framework adapts to what you have.
**Q: How do I update after publishing new papers?**
Run `/setup-extract` on the new paper, then update your experience file and bundles. Existing session files are not affected.
**Q: Can I use this with resume formats other than the included templates?**
Yes. The `.cls` files define the visual style. You can modify them or write your own. The skills generate content based on the template structure — update the `[GENERATE: ...]` and `[FIXED: ...]` markers in your template.
**Q: How long does the initial setup take?**
Depends on how many papers you have. Expect ~10 minutes per paper for extraction, then ~30 minutes for `/setup-build-kb` to synthesize everything. After that, each new JD takes about 15-20 minutes across all three sessions.
**Q: Can multiple people use the same kit?**
Each person needs their own clone with their own `config.md`, knowledge base, and templates. The framework itself is shared; the content is personal.
**Q: What Claude model should I use?**
The skills are designed for Claude's most capable models (Opus, Sonnet). Less capable models may skip steps or produce lower-quality output.
--- ---
## Contributing ## Contributing
Issues and PRs welcome. If you find a bug in the skill prompts, critique framework, or templates, please open an issue. Issues and PRs welcome. When contributing:
- Example files use the fictional Dr. Jordan Chen — keep examples in that persona
When contributing, keep in mind: - Reference docs should stay domain-agnostic
- The example files use the fictional Dr. Jordan Chen — keep examples in that persona
- Reference docs should stay domain-agnostic (no field-specific examples)
- Test skill changes against the example data before submitting - Test skill changes against the example data before submitting
--- ---
resume_builder/examples/example_cover_letter.pdf
BIN
View File
Binary file not shown.
resume_builder/examples/example_resume.pdf
BIN
View File
Binary file not shown.
resume_builder/examples/output/e2e_whitfield_proteineng_cover_letter.tex
+43
View File
@@ -0,0 +1,43 @@
% Example output — Dr. Jordan Chen cover letter for Whitfield University
% Generated by claude-resume-kit for demonstration purposes
% This is a fictional researcher; all data is fabricated.
\documentclass[11pt,a4paper,roman]{moderncv}
\usepackage[english]{babel}
\moderncvstyle{classic}
\moderncvcolor{green}
\usepackage[utf8]{inputenc}
\usepackage{ragged2e}
\usepackage[scale=0.79]{geometry}
\usepackage[version=4,arrows=pgf-filled]{mhchem}
\renewcommand*{\makeletterclosing}{\par\vspace{2ex}\closingname\par}
\name{Jordan}{Chen}
\address{Richland, WA 99354}
\phone[mobile]{+1 (555) 123-4567}
\email{jordan.chen@email.com}
\begin{document}
\recipient{To}{Hiring Committee\\Department of Biomedical Engineering\\Whitfield University\\Westbrook, MA 02481}
\date{\today}
\opening{Dear Members of the Hiring Committee,}
\makelettertitle
\begin{justify}
Your department's work at the intersection of structural biology and therapeutic design resonates with the research program I have built over the past three years: using machine learning to accelerate protein engineering decisions that would otherwise take months of experimental iteration. As a postdoctoral researcher developing ML-guided enzyme screening pipelines, I am excited to apply for the Assistant Professor position in Computational Protein Engineering (BME-2026-0042), where I would establish an independent group bridging data-driven protein design and biomolecular simulation.
At Lakewood University, I fine-tuned the ESM-2 protein language model on 45,000 experimental melting temperatures to screen 8,500 enzyme variants for industrial thermostability, compressing what would have been over a year of wet-lab work into 48 hours of computation. Five of our top seven candidates were confirmed experimentally by collaborators via differential scanning calorimetry. I then co-developed an open-source transfer learning framework that reduces labeled training data requirements by 60\%, now adopted by four external research groups. More recently, I extended our ML pipeline to predict enzyme tolerance across eight organic co-solvent systems, opening a pathway toward engineering biocatalysts for green chemistry. This research trajectory, from classical MD to ML-accelerated protein engineering, reflects the kind of program I would build at Whitfield.
My doctoral work at Westfield established the simulation foundations that make this ML approach rigorous rather than purely correlative. I developed enhanced sampling protocols that predict protein folding temperatures within 8 K of experiment, benchmarked four force fields for intrinsically disordered proteins, and calculated ligand binding free energies across three drug target families with sub-kcal/mol accuracy. I also built the curated thermostability database that directly enabled my postdoctoral ML work. Throughout, I mentored three graduate students and developed computational lab modules now used department-wide.
I would welcome the opportunity to contribute to Whitfield's strengths in biomaterials and therapeutic design. I could teach courses in computational biology, molecular modeling, and machine learning for biomedical applications. I look forward to discussing how my research program would complement your department's existing strengths and the collaborative opportunities available through your HPC infrastructure.
\end{justify}
\vspace{0.3cm}
{Sincerely,\\
Jordan Chen, Ph.D.\\
Postdoctoral Research Associate\\
Lakewood University}
\end{document}
resume_builder/examples/output/e2e_whitfield_proteineng_resume.tex
+154
View File
@@ -0,0 +1,154 @@
% Example output — Dr. Jordan Chen applying to Whitfield University
% Generated by claude-resume-kit for demonstration purposes
% This is a fictional researcher; all data is fabricated.
\documentclass{resume}
\usepackage{hyperref}
\usepackage{enumitem}
\usepackage{fontawesome}
\usepackage{tikz}
\usepackage{graphicx}
\hypersetup{
colorlinks = true,
linkcolor = [rgb]{0.9,0.4,0.4},
anchorcolor = [rgb]{0.9,0.4,0.4},
citecolor = [rgb]{0.4,0.4,0.4},
filecolor = [rgb]{0.4,0.4,0.4},
urlcolor = [rgb]{0.0,0.0,0.99},
}
\usepackage{xcolor}
\usepackage[version=4,arrows=pgf-filled]{mhchem}
\usepackage[includefoot,left=0.5in,top=0.5in,right=0.5in,bottom=0.2in,textwidth=7.5in,textheight=10.8in]{geometry}
\usepackage{fancyhdr}
\pagestyle{fancy}
\fancyhf{}
\renewcommand{\headrulewidth}{0pt}
\fancyfoot[R]{\hfill \thepage/\pageref{LastPage}}
\newcommand{\tab}[1]{\hspace{.2667\textwidth}\rlap{#1}}
\newcommand{\itab}[1]{\hspace{0em}\rlap{#1}}
\name{Jordan Chen, Ph.D.}
\address{jordan.chen@email.com \\ +1 (555) 123-4567}
\address{Richland, WA (Open to relocation to Westbrook, MA)}
\address{{Computational Protein Engineering $\vert$ ML-Guided Enzyme Design $\vert$ Biomolecular Simulation}}
\begin{document}
\vspace{-0.15cm}
%----------------------------------------------------------------------------------------
% SUMMARY
%----------------------------------------------------------------------------------------
\begin{rSection}{Summary}
Computational biologist with 8+ years combining \textbf{protein language models} and \textbf{molecular dynamics simulations} for enzyme engineering and drug discovery. Fine-tuned ESM-2 on 45K experimental stability measurements to screen 8,500 enzyme variants at 3,000$\times$ experimental throughput, with 5 hits confirmed by collaborators. Co-developed open-source transfer learning framework adopted by 4 external groups. 15 publications (7 first-author) in ACS Catalysis, J.\ Chem.\ Theory Comput., and J.\ Med.\ Chem.
\end{rSection}
\vspace{-0.15cm}
%----------------------------------------------------------------------------------------
% TECHNICAL SKILLS
%----------------------------------------------------------------------------------------
\begin{rSection}{Technical Skills}
\begin{skillgroup}{Molecular Simulation \& Modeling}
\skilldash{\textbf{GROMACS}, OpenMM, AMBER -- metadynamics, replica exchange MD, free energy perturbation}
\skilldash{AlphaFold2, Rosetta, AutoDock Vina -- protein structure prediction and molecular docking}
\skilldash{CHARMM36m, AMBER ff19SB, OPLS-AA/M -- force field benchmarking for disordered proteins}
\skilldash{Collective variable design, enhanced sampling protocol development, convergence analysis}
\end{skillgroup}
\begin{skillgroup}{Machine Learning \& Data Science}
\skilldash{\textbf{Protein language models} (ESM-2), graph neural networks, transfer learning, active learning}
\skilldash{\textbf{PyTorch}, scikit-learn, BioPython -- model fine-tuning, feature engineering, sequence analysis}
\skilldash{Regression, cross-validation, Spearman/RMSE benchmarking, dataset curation from public DBs}
\end{skillgroup}
\begin{skillgroup}{Programming \& HPC}
\skilldash{\textbf{Python}, Bash, SQL -- scientific computing, analysis pipelines, database management}
\skilldash{\textbf{SLURM}, Snakemake, Git, DVC -- HPC workflow automation and reproducible research}
\end{skillgroup}
\begin{skillgroup}{Analysis \& Visualization}
\skilldash{MDAnalysis, ProDy, PyMOL, matplotlib, seaborn -- trajectory analysis, publication figures}
\skilldash{PostgreSQL, pandas -- curated stability databases with automated quality filters for ML}
\end{skillgroup}
\begin{skillgroup}{Domain Expertise}
\skilldash{Protein engineering, enzyme thermostability, folding thermodynamics, drug discovery}
\skilldash{Intrinsically disordered proteins, ligand binding, biocatalysis, directed evolution}
\end{skillgroup}
\end{rSection}
\vspace{-0.15cm}
%----------------------------------------------------------------------------------------
% RESEARCH EXPERIENCE
%----------------------------------------------------------------------------------------
\begin{rSection}{Research Experience}
\begin{rSubsection}{ML-Accelerated Protein Engineering and Computational Enzyme Design}{\textcolor{black!60}{Aug 2023 -- Present}}{Postdoctoral Research Associate, Lakewood University}{}
\item Fine-tuned ESM-2 protein language model on 45K experimental melting temperatures, achieving 0.82 Spearman correlation and enabling 3,000$\times$ throughput screening of 8,500 enzyme variants for industrial thermostability.
\item Co-developed transfer learning framework from protein language models reducing labeled training data by 60\% across 5 enzyme families, released as open-source tool with 200+ GitHub stars.
\item Extended protein language model to predict enzyme solvent tolerance across 8 organic co-solvent systems, validating against 50-ns explicit-solvent MD for 80 enzyme variants and identifying 4 candidates for green chemistry.
\item Automated sequence-to-simulation pipeline using Snakemake workflow manager, reducing per-variant setup from 4 hours to 10 minutes and supporting 6 researchers across 3 active projects.
\end{rSubsection}
\begin{rSubsection}{Enhanced Sampling Methods for Protein Folding and Ligand Binding}{\textcolor{black!60}{Aug 2018 -- Jul 2023}}{Ph.D.\ Researcher, Westfield Institute of Technology}{}
\item Developed metadynamics-based enhanced sampling protocol for protein folding free energy landscapes, predicting folding temperatures within 8 K of experiment across 6 small proteins.
\item Calculated relative binding free energies for 40 congeneric ligand pairs via free energy perturbation, achieving 0.9 kcal/mol RMSE against experimental IC50 data across 3 drug target families.
\item Built curated protein thermostability database integrating 12,000 experimental melting temperatures from 3 public sources, with automated quality filters adopted by 8 lab members for ML training set construction.
\end{rSubsection}
\begin{rSubsection}{Computational Biophysics and Structural Analysis}{\textcolor{black!60}{May 2016 -- Jul 2018}}{Undergraduate Research Assistant, Eastgate University}{}
\item Performed homology modeling and 100-ns MD simulations of 4 mutant lysozyme variants, identifying destabilizing cavity mutations consistent with published experimental unfolding data.
\item Built Python analysis scripts for automated hydrogen bond occupancy tracking across 500-ns aggregate trajectories, adopted by 3 lab members for ongoing protein stability projects.
\end{rSubsection}
\end{rSection}
\vspace{-0.15cm}
%----------------------------------------------------------------------------------------
% EDUCATION
%----------------------------------------------------------------------------------------
\begin{rSection}{Education}
{Ph.D., Biomedical Engineering} \hfill {\textcolor{black!60}{Aug 2018 -- Jul 2023}}\\
{Westfield Institute of Technology}, Westfield, MA \hfill GPA: \textbf{3.92}/4.00
{B.S., Biochemistry (Honors)} \hfill {\textcolor{black!60}{Aug 2014 -- May 2018}}\\
{Eastgate University}, Portland, OR \hfill GPA: \textbf{3.87}/4.00
\end{rSection}
\vspace{-0.15cm}
%----------------------------------------------------------------------------------------
% SELECTED PUBLICATIONS
%----------------------------------------------------------------------------------------
\begin{rSection2}{Selected Publications (15 papers $\vert$ 280+ citations)}
\item \textbf{J.\ Chen}, R.\ Nakamura, S.\ Patel, K.\ Holmberg, M.\ Rivera. ``Deep Learning-Guided Screening of Thermostable Enzyme Variants for Industrial Biocatalysis.'' \textit{ACS Catalysis}, 2025.
\item \textbf{J.\ Chen}, M.\ Rivera, K.\ Holmberg. ``Transfer Learning from Protein Language Models for Low-Data Enzyme Property Prediction.'' \textit{Bioinformatics}, 2024.
\item \textbf{J.\ Chen}, L.\ Alvarez. ``Ligand Binding Free Energies via Enhanced-Sampling FEP for Three Drug Target Families.'' \textit{J.\ Med.\ Chem.}, 2023.
\item \textbf{J.\ Chen}, L.\ Alvarez. ``Metadynamics Protocol for Protein Folding Free Energy Landscapes.'' \textit{J.\ Chem.\ Theory Comput.}, 2022.
\item \textbf{J.\ Chen}, P.\ Kowalski, L.\ Alvarez. ``Force Field Benchmarking for Intrinsically Disordered Protein Ensembles.'' \textit{J.\ Chem.\ Theory Comput.}, 2021.
\end{rSection2}
\vspace{-0.15cm}
%----------------------------------------------------------------------------------------
% HONORS & AWARDS
%----------------------------------------------------------------------------------------
\begin{rSection2}{Honors \& Awards}
\item \textbf{NSF Graduate Research Fellowship}, National Science Foundation (2019). Three-year fellowship supporting doctoral research in computational protein engineering.
\item \textbf{Best Oral Presentation}, Westfield Biophysics Symposium (2022). Enhanced sampling methods for protein folding thermodynamics.
\item \textbf{Dean's Teaching Award}, Westfield Institute of Technology (2021). Outstanding TA in computational biology.
\end{rSection2}
\vspace{-0.1cm}
\begin{center}
\vspace{0.15cm}
\textit{Authorized to work in the United States}
\end{center}
\end{document}