niw-skill-suite
Open-source AI skills and plugins for EB-1 and EB-2 NIW immigration petitions — from evidence organization to RFE response.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/superme-ai-skills--niw-skill-suite)HTML
<a href="/listings/superme-ai-skills--niw-skill-suite"><img src="/api/listings/superme-ai-skills--niw-skill-suite/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Requirements
- Claude Code, claude.ai with Skills enabled, or Codex with local skills support - Code execution enabled when using tools that require it
Installation
There are three ways to install: **plugins** for Claude Code, **individual `.skill` uploads** for claude.ai, and **extracted skill folders** for Codex.
Option A — Plugin (Claude Code / CLI)
Plugins bundle all skills for a petition type into a single file. Install via double-click or terminal:
# Clone the repo git clone https://github.com/VeraSuperHub/vera-eb-suite.git cd vera-eb-suite # Install the plugin(s) you need claude plugin install vera-niw.plugin claude plugin install vera-eb1.plugin
If the `.plugin` file extension is not recognized on your system, rename it to `.zip` before installing:
cp vera-niw.plugin vera-niw.zip claude plugin install vera-niw.zip
Option B — Individual Skills (claude.ai)
For use on [claude.ai](https://claude.ai), install skills one at a time:
1. Download the `.skill` file(s) you need from [`vera-niw-skillset/`](vera-niw-skillset/) or [`vera-eb1-skillset/`](vera-eb1-skillset/) 2. Go to [Settings → Capabilities](https://claude.ai/settings/capabilities) 3. Scroll to the **Skills** section 4. Click **"Upload skill"** 5. Upload the `.skill` file and toggle it **on**
Claude will automatically invoke the skill when your request matches its description — no manual activation needed.
**Install one skill at a time.** For the full NIW pipeline, install all 8. For EB-1, install all 11.
Option C — Individual Skills (Codex)
For Codex, each `.skill` file is a zip archive containing a standard skill folder with `SKILL.md` plus its supporting `references/`, `docs/`, `evals/`, and schemas. Extract the archives into your Codex skills directory:
# Install all petition skills into the default Codex skills directory mkdir -p ~/.codex/skills for f in vera-niw-skillset/*.skill vera-eb1-skillset/*.skill; do unzip -oq "$f" -d ~/.codex/skills done
After extraction, Codex can discover the skills from `~/.codex/skills/<skill-name>/SKILL.md`.
Google Scholar Tool Setup
The `GoogleScholar/` directory contains a Python scraper for extracting citation metrics. You can run it locally or via Google Colab:
cd GoogleScholar
pip install requests beautifulsoup4 pandas numpy
python -c "from scholar import get_profile; print(get_profile('YOUR_SCHOLAR_ID'))"Or open `scholar_colab_demo.ipynb` in [Google Colab](https://colab.research.google.com) for interactive use.
Recommended Workflow
Start with **Evaluate** to get your evidence-readiness summary — a structured map of strengths, gaps, and risk areas, intended for human review. If the readiness summary indicates the evidence base looks workable for review with an attorney, proceed through the pipeline in order:
**NIW:**
Evaluate → Endeavor → Pillar (×3) → Recommendation (×N) → Assemble → PL Review
Each skill's output is designed as input for the next skill in the pipeline. The Evaluate JSON feeds into Endeavor, Endeavor's pillar seeds feed into Pillar, and all outputs converge in Assemble.
**EB-1:**
Evaluate → Criterion Skills (select based on your evidence) → Recommendation (×N) + Final Merits → Assemble → PL Review
The Evaluate skill produces pathway framing (EB-1A vs EB-1B) and identifies which criteria your evidence appears to support — both as drafting inputs for human and attorney review. Run the relevant criterion skills, then generate recommendation-letter drafts and the final-merits argument support before assembling the petition draft.
---
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.
README.md · README · 378 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
- # Vera EB Suite
- > Hi, I'm **Vera** — a silicon-based rabbit who documents the open-source skills Veronica created.
- >
- > Veronica has a PhD in Quantitative Sciences, 10+ years across quantitative research, AI, and clinical trials, with publications in psychometrics and human-AI collaboration. She also went through the NIW process herself. She created this suite to systematize the parts of evidence-building and petition preparation that can be decomposed, documented, and reviewed. I help structure the workflows. She reviews, tests, and decides what ships.
- >
- > Everything in this repo is what can be made explicit: evidence organization, gap spotting, document drafting support, and review workflows. What the suite cannot do is assess whether your specific case will be approved, provide legal advice, or replace an experienced immigration attorney. That remains a human and legal judgment.
- **Open-source Claude Code and Codex-compatible skills for EB-1 and EB-2 NIW evidence-building and petition-preparation support — from evidence review and case organization to drafting workflows, recommendation-letter support, pre-filing review, and RFE response preparation.**
- Each skill encodes structured reasoning patterns derived from public USCIS materials, AAO decisions, policy guidance, and evidence-organization workflows. The repo ships Claude Code plugin bundles plus plain `SKILL.md` folders inside `.skill` archives that can also be installed into Codex.
- > _Across [VeraSuperHub](https://github.com/VeraSuperHub), Vera structures execution; humans own judgment._
- > **Why this exists:** Immigration petitions are high-stakes, and information asymmetry can make the process harder than it needs to be. Many parts of evidence-building and petition preparation follow patterns that can be made explicit: organizing exhibits, identifying gaps, mapping evidence to criteria, drafting structured narratives, and stress-testing a petition before filing. This project decomposes those repeatable parts into modular, testable, improvable skills — while leaving legal strategy, approval assessment, and final judgment to qualified human professionals.
- ---
- ## The Pipelines
- ### NIW Pipeline (EB-2 National Interest Waiver)
- ```
- ┌──────────────┐ ┌──────────────┐ ┌──────────────┐
- │ 1. EVALUATE │────▶│ 2. ENDEAVOR │────▶│ 3. PILLAR │
- │ Evidence- │ │ Endeavor │ │ ×3 runs │
- │ readiness │ │ statement │ │ (one per │
- │ summary │ │ + 3 pillar │ │ pillar) │
- └──────────────┘ │ seeds │ └──────┬───────┘
- └──────────────┘ │
- ▼
- ┌─────────────┐ ┌─────────────┐ ┌──────────────┐
- │ 7. RFE │ │ 6. PL │◀────│ 5. ASSEMBLE │
- │ RESPONSE │ │ REVIEW │ │ Review- │
- │ DRAFTING │ │ Adversarial │ │ ready │
- │ SUPPORT │ │ pre-filing │ │ draft .docx │
- └─────────────┘ │ check │ └──────┬───────┘
- └─────────────┘ │
- ▲ ┌─────┴────────┐
- └──────────────│ 4. RECOMMEND │
- │ Reference │
- │ letters │
- └──────────────┘
- ```
- Entrepreneur cases route through **vera-niw-entrepreneur** before entering the standard pipeline at Step 2.
- ### EB-1 Pipeline (Extraordinary Ability / Outstanding Researcher)
- > **STEM focus:** The criterion skills below cover the criteria most commonly used in STEM petitions. This is not the full set of EB-1 criteria — criteria such as awards (Crit. 1), membership (Crit. 2), high salary (Crit. 9), and commercial success (Crit. 10) are not yet included. For EB-1A, petitioners must meet at least 3 of the 10 criteria; for EB-1B, petitioners must meet at least 2 of the 6 criteria. Use the criterion skills that match your evidence.
- ```
- ┌──────────────┐ ┌──────────────────────────────────────┐
- │ 1. EVALUATE │────▶│ 2. CRITERION SKILLS │
- │ Evidence- │ │ ┌────────────┐ ┌────────────────┐ │
- │ readiness │ │ │ AUTHORSHIP │ │ ORIGINAL │ │
- │ summary + │ │ │ (Crit. 6) │ │ CONTRIBUTIONS │ │
- │ pathway │ │ └────────────┘ │ (Crit. 5) │ │
- │ framing │ │ ┌────────────┐ └────────────────┘ │
- │ (EB-1A / │ │ │ JUDGING │ ┌────────────────┐ │
- │ EB-1B) │ │ │ (Crit. 4) │ │ CRITICAL ROLE │ │
- └──────────────┘ │ └────────────┘ │ (Crit. 8) │ │
- │ ┌────────────┐ └────────────────┘ │
- │ │ PUBLISHED │ │
- │ │ MATERIAL │ │
- │ │ (Crit. 3) │ │
- │ └────────────┘ │
- └──────────────────┬───────────────────┘
- ▼
- ┌─────────────┐ ┌─────────────┐ ┌──────────────┐
- │ 6. RFE │ │ 5. PL │◀─│ 4. ASSEMBLE │
- │ RESPONSE │ │ REVIEW │ │ Review- │
- │ DRAFTING │ │ Adversarial │ │ ready │
- │ SUPPORT │ │ pre-filing │ │ draft .docx │
- └─────────────┘ │ check │ └──────┬───────┘
- └─────────────┘ │
- ▲ ┌─────┴────────┐
- └───────────│ 3. RECOMMEND │
- │ + FINAL │
- │ MERITS │
- │ DRAFTING │
- └──────────────┘
- ```
- ---
- ## Skills
- ### NIW ([`vera-niw.plugin`](vera-niw.plugin) / [`vera-niw-skillset/`](vera-niw-skillset/))
- | # | Skill | What It Does |
- |---|-------|-------------|
- | 1 | `vera-niw-evaluate` | Reviews the petitioner's profile, maps available evidence to NIW requirements, identifies strengths, gaps, and risk areas, and produces an evidence-readiness summary for human review |
- | 2 | `vera-niw-endeavor` | Drafts a proposed endeavor statement and related framing options based on the petitioner's field, evidence, and intended contribution, for human review and revision |
- | 3 | `vera-niw-pillar` | Drafts structured petition-letter sections for the three NIW prongs: substantial merit and national importance, well-positioned, and balance of equities. Designed for human review and revision |
- | 4 | `vera-niw-recommendation` | Drafts recommendation-letter support materials with recommender-specific framing, evidence mapping, and nonredundant emphasis areas for human review |
- | 5 | `vera-niw-assemble` | Assembles a review-ready petition-support package — petition-letter draft, exhibit list, and supporting-document structure — with cross-reference checks |
- | 6 | `vera-niw-pl-review` | Runs an adversarial pre-filing review using public AAO reasoning patterns and common evidence weaknesses to identify gaps, ambiguity, and potential RFE triggers |
- | 7 | `vera-niw-rfe-response` | Drafts a structured point-by-point RFE response framework that organizes each USCIS finding, relevant evidence, updated metrics, and potential new exhibits for human review |
- | 8 | `vera-niw-entrepreneur` | Reviews entrepreneur/founder NIW evidence using the USCIS Policy Manual's entrepreneur-specific framework and identifies evidence gaps, narrative risks, and documentation needs |
- > **Got a weak research profile?** If `vera-niw-evaluate` or `vera-eb1-evaluate` flags thin publications or citation impact, see [**ai-research-pipeline**](https://github.com/VeraSuperHub/ai-research-pipeline) and [**stat-research-pipeline**](https://github.com/VeraSuperHub/stat-research-pipeline) — sister skill suites that structure the execution layer of an AI/statistical research workflow (diagnostics, candidate analyses, manuscript-section drafting, review checkpoints) for human-led research production.
- ### EB-1 ([`vera-eb1.plugin`](vera-eb1.plugin) / [`vera-eb1-skillset/`](vera-eb1-skillset/))
- | # | Skill | What It Does |
- |---|-------|-------------|
- | 1 | `vera-eb1-evaluate` | Reviews evidence for EB-1A and EB-1B pathways, maps available materials to relevant criteria, identifies strengths, gaps, and risk areas, and produces an evidence-readiness summary for human review |
- | 2 | `vera-eb1-authorship` | Maps scholarly authorship evidence to the relevant EB-1 criterion, including publication venues, citation context, authorship role, and evidence gaps for human review |
- | 3 | `vera-eb1-original-contributions` | Drafts support for organizing original-contribution evidence, including contribution framing, adoption or impact signals, before/after context, and documentation gaps for human review |
- | 4 | `vera-eb1-judging` | Organizes judging evidence, such as peer review, panels, editorial service, and evaluation roles, and maps it to the relevant EB-1 criterion for human review |
- | 5 | `vera-eb1-critical-role` | Organizes evidence for leading or critical roles, including role context, organizational distinction, scope of responsibility, and impact documentation for human review |
- | 6 | `vera-eb1-published-material` | Organizes published-material evidence about the petitioner, including source credibility, media context, relevance, and documentation gaps for human review |
- | 7 | `vera-eb1-recommendation` | Drafts recommendation-letter support materials with recommender-specific framing, evidence mapping, and nonredundant emphasis areas for human review |
- | 8 | `vera-eb1-final-merits` | Drafts final-merits argument support by organizing sustained-acclaim evidence, criteria-level outputs, impact signals, and narrative risks for human review |
- | 9 | `vera-eb1-assemble` | Assembles a review-ready EB-1 petition-letter draft and supporting evidence structure as a formatted .docx |
- | 10 | `vera-eb1-pl-review` | Runs an adversarial pre-filing review using the Kazarian two-step framework, public AAO reasoning patterns, and common evidence weaknesses to identify gaps, ambiguity, and potential RFE triggers |
- | 11 | `vera-eb1-rfe-response` | Drafts a structured point-by-point EB-1 RFE response framework that organizes each USCIS finding, relevant evidence, updated metrics, and potential new exhibits for human review |
- **Total: 19 skills across both petition categories.**
- > **Got a weak research profile?** If `vera-niw-evaluate` or `vera-eb1-evaluate` flags thin publications or citation impact, see [**ai-research-pipeline**](https://github.com/VeraSuperHub/ai-research-pipeline) and [**stat-research-pipeline**](https://github.com/VeraSuperHub/stat-research-pipeline) — sister skill suites that structure the execution layer of an AI/statistical research workflow (diagnostics, candidate analyses, manuscript-section drafting, review checkpoints) for human-led research production.
- ---
- ## Tools
- In addition to skills, this suite includes standalone tools that feed data into the pipeline:
- | Tool | What It Does | Used By |
- |---|---|---|
- | [`GoogleScholar`](GoogleScholar/) | Extracts citation metrics, publication lists, and h-index from Google Scholar (Python + Colab notebook) | `vera-niw-assemble` (Section 3: Academic Credentials) |
- ---
- ## Quick Start
- ### Requirements
- - Claude Code, claude.ai with Skills enabled, or Codex with local skills support
- - Code execution enabled when using tools that require it
- ### Installation
- There are three ways to install: **plugins** for Claude Code, **individual `.skill` uploads** for claude.ai, and **extracted skill folders** for Codex.
- #### Option A — Plugin (Claude Code / CLI)
- Plugins bundle all skills for a petition type into a single file. Install via double-click or terminal:
- ```bash
- # Clone the repo
- git clone https://github.com/VeraSuperHub/vera-eb-suite.git
- cd vera-eb-suite
- # Install the plugin(s) you need
- claude plugin install vera-niw.plugin
- claude plugin install vera-eb1.plugin
- ```
- If the `.plugin` file extension is not recognized on your system, rename it to `.zip` before installing:
- ```bash
- cp vera-niw.plugin vera-niw.zip
- claude plugin install vera-niw.zip
- ```
- #### Option B — Individual Skills (claude.ai)
- For use on [claude.ai](https://claude.ai), install skills one at a time:
- 1. Download the `.skill` file(s) you need from [`vera-niw-skillset/`](vera-niw-skillset/) or [`vera-eb1-skillset/`](vera-eb1-skillset/)
- 2. Go to [Settings → Capabilities](https://claude.ai/settings/capabilities)
- 3. Scroll to the **Skills** section
- 4. Click **"Upload skill"**
- 5. Upload the `.skill` file and toggle it **on**
- Claude will automatically invoke the skill when your request matches its description — no manual activation needed.
- **Install one skill at a time.** For the full NIW pipeline, install all 8. For EB-1, install all 11.
- #### Option C — Individual Skills (Codex)
- For Codex, each `.skill` file is a zip archive containing a standard skill folder with `SKILL.md` plus its supporting `references/`, `docs/`, `evals/`, and schemas. Extract the archives into your Codex skills directory:
- ```bash
- # Install all petition skills into the default Codex skills directory
- mkdir -p ~/.codex/skills
- for f in vera-niw-skillset/*.skill vera-eb1-skillset/*.skill; do
- unzip -oq "$f" -d ~/.codex/skills
- done
- ```
- After extraction, Codex can discover the skills from `~/.codex/skills/<skill-name>/SKILL.md`.
- ### Google Scholar Tool Setup
- The `GoogleScholar/` directory contains a Python scraper for extracting citation metrics. You can run it locally or via Google Colab:
- ```bash
- cd GoogleScholar
- pip install requests beautifulsoup4 pandas numpy
- python -c "from scholar import get_profile; print(get_profile('YOUR_SCHOLAR_ID'))"
- ```
- Or open `scholar_colab_demo.ipynb` in [Google Colab](https://colab.research.google.com) for interactive use.
- ### Recommended Workflow
- Start with **Evaluate** to get your evidence-readiness summary — a structured map of strengths, gaps, and risk areas, intended for human review. If the readiness summary indicates the evidence base looks workable for review with an attorney, proceed through the pipeline in order:
- **NIW:**
- ```
- Evaluate → Endeavor → Pillar (×3) → Recommendation (×N) → Assemble → PL Review
- ```
- Each skill's output is designed as input for the next skill in the pipeline. The Evaluate JSON feeds into Endeavor, Endeavor's pillar seeds feed into Pillar, and all outputs converge in Assemble.
- **EB-1:**
- ```
- Evaluate → Criterion Skills (select based on your evidence) → Recommendation (×N) + Final Merits → Assemble → PL Review
- ```
- The Evaluate skill produces pathway framing (EB-1A vs EB-1B) and identifies which criteria your evidence appears to support — both as drafting inputs for human and attorney review. Run the relevant criterion skills, then generate recommendation-letter drafts and the final-merits argument support before assembling the petition draft.
- ---
- ## Usage Examples
- **Evaluate — How strong is my evidence?**
- ```
- I'm a senior data scientist at a Fortune 500 company with 5 years of experience.
- I have 3 publications (12 citations total), 2 patents pending, and my fraud
- detection system processes 2M+ transactions daily. Review my NIW evidence
- readiness.
- ```
- **Endeavor — Define the proposed endeavor:**
- ```
- I just completed vera-niw-evaluate and got a WORKABLE_EVIDENCE readiness tier.
- Here's my JSON output:
- [paste evaluate output]
- Help me draft my proposed endeavor.
- ```
- **Pillar — Write petition content:**
- ```
- Here's my endeavor statement and three pillar definitions from NIW_Endeavor:
- [paste endeavor output]
- Write the petition content for Pillar 1.
- ```
- **PL Review — Adversarial pre-filing check:**
- ```
- Run a structured adversarial review of my completed petition letter using
- public AAO reasoning patterns and common evidence weaknesses. Surface every
- gap, ambiguity, or pattern that would commonly trigger an RFE.
- [paste petition letter]
- ```
- **RFE Response — Drafting support:**
- ```
- I received this RFE on my NIW petition. Here's the RFE notice and my
- original petition letter. Draft a structured point-by-point response
- framework that maps each USCIS finding to relevant evidence and potential
- new exhibits, for human and attorney review.
- [paste RFE notice]
- [paste original petition]
- ```
- ---
- ## How It Works
- Each skill encodes structured reasoning patterns as a decomposable workflow:
- ```
- Public materials (USCIS Policy Manual, AAO decisions, policy guidance)
- ↓
- Decompose into evidence-organization rubrics & checklists
- ↓
- Encode as structured skill instructions (SKILL.md)
- ↓
- Add reference materials (rubrics, schemas, examples)
- ↓
- Validate against test cases (evals/)
- ```
- The key insight: many parts of evidence preparation are pattern-based. Public decisions, policy guidance, and petition workflows reveal recurring structures — what evidence tends to be persuasive, where gaps commonly appear, how narratives are organized, and how weaknesses are stress-tested before filing. These repeatable components can be decomposed, encoded, validated, and improved by the community — while case-specific legal judgment remains a human professional responsibility.
- ---
- ## Repo Structure
- ```
- vera-eb-suite/
- ├── README.md
- ├── LICENSE (GPL-3.0)
- ├── DISCLAIMER.md (legal disclaimer)
- ├── CONTRIBUTING.md (contribution guidelines)
- ├── CHANGELOG.md (version history)
- │
- ├── vera-niw-skillset/ ← NIW skills (8 .skill files)
- │ ├── vera-niw-evaluate.skill
- │ ├── vera-niw-endeavor.skill
- │ ├── vera-niw-pillar.skill
- │ ├── vera-niw-recommendation.skill
- │ ├── vera-niw-assemble.skill
- │ ├── vera-niw-pl-review.skill
- │ ├── vera-niw-rfe-response.skill
- │ └── vera-niw-entrepreneur.skill
- │
- ├── vera-eb1-skillset/ ← EB-1 skills (11 .skill files)
- │ ├── vera-eb1-evaluate.skill
- │ ├── vera-eb1-authorship.skill
- │ ├── vera-eb1-original-contributions.skill
- │ ├── vera-eb1-judging.skill
- │ ├── vera-eb1-critical-role.skill
- │ ├── vera-eb1-published-material.skill
- │ ├── vera-eb1-recommendation.skill
- │ ├── vera-eb1-final-merits.skill
- │ ├── vera-eb1-assemble.skill
- │ ├── vera-eb1-pl-review.skill
- │ └── vera-eb1-rfe-response.skill
- │
- ├── vera-niw.plugin ← Bundled NIW plugin
- ├── vera-eb1.plugin ← Bundled EB-1 plugin
- │
- └── GoogleScholar/ ← Citation data tool
- ├── scholar.py
- ├── scholar_colab_demo.ipynb
- └── requirements.txt
- ```
- ---
- ## FAQ
- **Is this legal advice?**
- No. See [DISCLAIMER.md](DISCLAIMER.md). These tools provide informational guidance only. Always consult a qualified immigration attorney for your specific case.
- **Do I need to be technical?**
- No. Claude users can install plugins or upload `.skill` files. Codex users can extract the same `.skill` archives into their local skills directory.
- **Will this assess my approval chances?**
- No. These skills do not predict or assess approval likelihood — that is a case-specific legal judgment that belongs to a qualified immigration attorney. The skills help you identify evidence gaps, organize exhibits, draft review-ready petition sections, and stress-test a petition before filing.
- **How is this different from ChatGPT prompts for NIW?**
- Generic prompts produce generic output. Each skill here encodes structured reasoning patterns — evidence-organization rubrics, common-weakness checks, field-specific framing, USCIS-language conventions — derived from public AAO decisions and the USCIS Policy Manual. The skills include explicit rubrics, schemas, failure-pattern checks, and evidence-mapping steps that make the output easier to review, test, and improve. They are drafting and review-support tools, not legal decision tools.
- **Can I use this with Codex, GPT-4, or other models?**
- Yes. The skills are structured as model-agnostic `SKILL.md` instruction folders. Claude-specific plugin files are provided for Claude Code, while Codex can use the extracted skill folders directly. Output quality may vary by model and context window.
- **Can I contribute?**
- Yes. See [CONTRIBUTING.md](CONTRIBUTING.md).
- ---
- ## Contributing
- Contributions welcome. See [CONTRIBUTING.md](CONTRIBUTING.md) for details. Especially valuable:
- - **Bug reports** — Skill produced incorrect or misleading guidance? Open an issue.
- - **Test cases** — Real RFE patterns or edge cases for `evals/`. Anonymize all personal information.
- - **Reference materials** — New AAO decisions, policy updates, or adjudication trend data.
- - **Skill improvements** — Better rubrics, additional failure patterns, improved prompts.
- ---
- ## Disclaimer
- These tools are for **informational and educational purposes only**. They do not constitute legal advice and do not create an attorney-client relationship. See [DISCLAIMER.md](DISCLAIMER.md) for full terms.
- ---
- ## License
- GPL-3.0 License. See [LICENSE](LICENSE).
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.