Skip to content
Back to search

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.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/superme-ai-skills--niw-skill-suite/badge.svg)](/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

i

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)

  1. # Vera EB Suite
  2. > Hi, I'm **Vera** — a silicon-based rabbit who documents the open-source skills Veronica created.
  3. >
  4. > 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.
  5. >
  6. > 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.
  7. **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.**
  8. 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.
  9. > _Across [VeraSuperHub](https://github.com/VeraSuperHub), Vera structures execution; humans own judgment._
  10. > **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.
  11. ---
  12. ## The Pipelines
  13. ### NIW Pipeline (EB-2 National Interest Waiver)
  14. ```
  15. ┌──────────────┐ ┌──────────────┐ ┌──────────────┐
  16. │ 1. EVALUATE │────▶│ 2. ENDEAVOR │────▶│ 3. PILLAR │
  17. │ Evidence- │ │ Endeavor │ │ ×3 runs │
  18. │ readiness │ │ statement │ │ (one per │
  19. │ summary │ │ + 3 pillar │ │ pillar) │
  20. └──────────────┘ │ seeds │ └──────┬───────┘
  21. └──────────────┘ │
  22. ▼
  23. ┌─────────────┐ ┌─────────────┐ ┌──────────────┐
  24. │ 7. RFE │ │ 6. PL │◀────│ 5. ASSEMBLE │
  25. │ RESPONSE │ │ REVIEW │ │ Review- │
  26. │ DRAFTING │ │ Adversarial │ │ ready │
  27. │ SUPPORT │ │ pre-filing │ │ draft .docx │
  28. └─────────────┘ │ check │ └──────┬───────┘
  29. └─────────────┘ │
  30. ▲ ┌─────┴────────┐
  31. └──────────────│ 4. RECOMMEND │
  32. │ Reference │
  33. │ letters │
  34. └──────────────┘
  35. ```
  36. Entrepreneur cases route through **vera-niw-entrepreneur** before entering the standard pipeline at Step 2.
  37. ### EB-1 Pipeline (Extraordinary Ability / Outstanding Researcher)
  38. > **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.
  39. ```
  40. ┌──────────────┐ ┌──────────────────────────────────────┐
  41. │ 1. EVALUATE │────▶│ 2. CRITERION SKILLS │
  42. │ Evidence- │ │ ┌────────────┐ ┌────────────────┐ │
  43. │ readiness │ │ │ AUTHORSHIP │ │ ORIGINAL │ │
  44. │ summary + │ │ │ (Crit. 6) │ │ CONTRIBUTIONS │ │
  45. │ pathway │ │ └────────────┘ │ (Crit. 5) │ │
  46. │ framing │ │ ┌────────────┐ └────────────────┘ │
  47. │ (EB-1A / │ │ │ JUDGING │ ┌────────────────┐ │
  48. │ EB-1B) │ │ │ (Crit. 4) │ │ CRITICAL ROLE │ │
  49. └──────────────┘ │ └────────────┘ │ (Crit. 8) │ │
  50. │ ┌────────────┐ └────────────────┘ │
  51. │ │ PUBLISHED │ │
  52. │ │ MATERIAL │ │
  53. │ │ (Crit. 3) │ │
  54. │ └────────────┘ │
  55. └──────────────────┬───────────────────┘
  56. ▼
  57. ┌─────────────┐ ┌─────────────┐ ┌──────────────┐
  58. │ 6. RFE │ │ 5. PL │◀─│ 4. ASSEMBLE │
  59. │ RESPONSE │ │ REVIEW │ │ Review- │
  60. │ DRAFTING │ │ Adversarial │ │ ready │
  61. │ SUPPORT │ │ pre-filing │ │ draft .docx │
  62. └─────────────┘ │ check │ └──────┬───────┘
  63. └─────────────┘ │
  64. ▲ ┌─────┴────────┐
  65. └───────────│ 3. RECOMMEND │
  66. │ + FINAL │
  67. │ MERITS │
  68. │ DRAFTING │
  69. └──────────────┘
  70. ```
  71. ---
  72. ## Skills
  73. ### NIW ([`vera-niw.plugin`](vera-niw.plugin) / [`vera-niw-skillset/`](vera-niw-skillset/))
  74. | # | Skill | What It Does |
  75. |---|-------|-------------|
  76. | 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 |
  77. | 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 |
  78. | 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 |
  79. | 4 | `vera-niw-recommendation` | Drafts recommendation-letter support materials with recommender-specific framing, evidence mapping, and nonredundant emphasis areas for human review |
  80. | 5 | `vera-niw-assemble` | Assembles a review-ready petition-support package — petition-letter draft, exhibit list, and supporting-document structure — with cross-reference checks |
  81. | 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 |
  82. | 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 |
  83. | 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 |
  84. > **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.
  85. ### EB-1 ([`vera-eb1.plugin`](vera-eb1.plugin) / [`vera-eb1-skillset/`](vera-eb1-skillset/))
  86. | # | Skill | What It Does |
  87. |---|-------|-------------|
  88. | 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 |
  89. | 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 |
  90. | 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 |
  91. | 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 |
  92. | 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 |
  93. | 6 | `vera-eb1-published-material` | Organizes published-material evidence about the petitioner, including source credibility, media context, relevance, and documentation gaps for human review |
  94. | 7 | `vera-eb1-recommendation` | Drafts recommendation-letter support materials with recommender-specific framing, evidence mapping, and nonredundant emphasis areas for human review |
  95. | 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 |
  96. | 9 | `vera-eb1-assemble` | Assembles a review-ready EB-1 petition-letter draft and supporting evidence structure as a formatted .docx |
  97. | 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 |
  98. | 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 |
  99. **Total: 19 skills across both petition categories.**
  100. > **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.
  101. ---
  102. ## Tools
  103. In addition to skills, this suite includes standalone tools that feed data into the pipeline:
  104. | Tool | What It Does | Used By |
  105. |---|---|---|
  106. | [`GoogleScholar`](GoogleScholar/) | Extracts citation metrics, publication lists, and h-index from Google Scholar (Python + Colab notebook) | `vera-niw-assemble` (Section 3: Academic Credentials) |
  107. ---
  108. ## Quick Start
  109. ### Requirements
  110. - Claude Code, claude.ai with Skills enabled, or Codex with local skills support
  111. - Code execution enabled when using tools that require it
  112. ### Installation
  113. There are three ways to install: **plugins** for Claude Code, **individual `.skill` uploads** for claude.ai, and **extracted skill folders** for Codex.
  114. #### Option A — Plugin (Claude Code / CLI)
  115. Plugins bundle all skills for a petition type into a single file. Install via double-click or terminal:
  116. ```bash
  117. # Clone the repo
  118. git clone https://github.com/VeraSuperHub/vera-eb-suite.git
  119. cd vera-eb-suite
  120. # Install the plugin(s) you need
  121. claude plugin install vera-niw.plugin
  122. claude plugin install vera-eb1.plugin
  123. ```
  124. If the `.plugin` file extension is not recognized on your system, rename it to `.zip` before installing:
  125. ```bash
  126. cp vera-niw.plugin vera-niw.zip
  127. claude plugin install vera-niw.zip
  128. ```
  129. #### Option B — Individual Skills (claude.ai)
  130. For use on [claude.ai](https://claude.ai), install skills one at a time:
  131. 1. Download the `.skill` file(s) you need from [`vera-niw-skillset/`](vera-niw-skillset/) or [`vera-eb1-skillset/`](vera-eb1-skillset/)
  132. 2. Go to [Settings → Capabilities](https://claude.ai/settings/capabilities)
  133. 3. Scroll to the **Skills** section
  134. 4. Click **"Upload skill"**
  135. 5. Upload the `.skill` file and toggle it **on**
  136. Claude will automatically invoke the skill when your request matches its description — no manual activation needed.
  137. **Install one skill at a time.** For the full NIW pipeline, install all 8. For EB-1, install all 11.
  138. #### Option C — Individual Skills (Codex)
  139. 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:
  140. ```bash
  141. # Install all petition skills into the default Codex skills directory
  142. mkdir -p ~/.codex/skills
  143. for f in vera-niw-skillset/*.skill vera-eb1-skillset/*.skill; do
  144. unzip -oq "$f" -d ~/.codex/skills
  145. done
  146. ```
  147. After extraction, Codex can discover the skills from `~/.codex/skills/<skill-name>/SKILL.md`.
  148. ### Google Scholar Tool Setup
  149. The `GoogleScholar/` directory contains a Python scraper for extracting citation metrics. You can run it locally or via Google Colab:
  150. ```bash
  151. cd GoogleScholar
  152. pip install requests beautifulsoup4 pandas numpy
  153. python -c "from scholar import get_profile; print(get_profile('YOUR_SCHOLAR_ID'))"
  154. ```
  155. Or open `scholar_colab_demo.ipynb` in [Google Colab](https://colab.research.google.com) for interactive use.
  156. ### Recommended Workflow
  157. 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:
  158. **NIW:**
  159. ```
  160. Evaluate → Endeavor → Pillar (×3) → Recommendation (×N) → Assemble → PL Review
  161. ```
  162. 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.
  163. **EB-1:**
  164. ```
  165. Evaluate → Criterion Skills (select based on your evidence) → Recommendation (×N) + Final Merits → Assemble → PL Review
  166. ```
  167. 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.
  168. ---
  169. ## Usage Examples
  170. **Evaluate — How strong is my evidence?**
  171. ```
  172. I'm a senior data scientist at a Fortune 500 company with 5 years of experience.
  173. I have 3 publications (12 citations total), 2 patents pending, and my fraud
  174. detection system processes 2M+ transactions daily. Review my NIW evidence
  175. readiness.
  176. ```
  177. **Endeavor — Define the proposed endeavor:**
  178. ```
  179. I just completed vera-niw-evaluate and got a WORKABLE_EVIDENCE readiness tier.
  180. Here's my JSON output:
  181. [paste evaluate output]
  182. Help me draft my proposed endeavor.
  183. ```
  184. **Pillar — Write petition content:**
  185. ```
  186. Here's my endeavor statement and three pillar definitions from NIW_Endeavor:
  187. [paste endeavor output]
  188. Write the petition content for Pillar 1.
  189. ```
  190. **PL Review — Adversarial pre-filing check:**
  191. ```
  192. Run a structured adversarial review of my completed petition letter using
  193. public AAO reasoning patterns and common evidence weaknesses. Surface every
  194. gap, ambiguity, or pattern that would commonly trigger an RFE.
  195. [paste petition letter]
  196. ```
  197. **RFE Response — Drafting support:**
  198. ```
  199. I received this RFE on my NIW petition. Here's the RFE notice and my
  200. original petition letter. Draft a structured point-by-point response
  201. framework that maps each USCIS finding to relevant evidence and potential
  202. new exhibits, for human and attorney review.
  203. [paste RFE notice]
  204. [paste original petition]
  205. ```
  206. ---
  207. ## How It Works
  208. Each skill encodes structured reasoning patterns as a decomposable workflow:
  209. ```
  210. Public materials (USCIS Policy Manual, AAO decisions, policy guidance)
  211. ↓
  212. Decompose into evidence-organization rubrics & checklists
  213. ↓
  214. Encode as structured skill instructions (SKILL.md)
  215. ↓
  216. Add reference materials (rubrics, schemas, examples)
  217. ↓
  218. Validate against test cases (evals/)
  219. ```
  220. 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.
  221. ---
  222. ## Repo Structure
  223. ```
  224. vera-eb-suite/
  225. ├── README.md
  226. ├── LICENSE (GPL-3.0)
  227. ├── DISCLAIMER.md (legal disclaimer)
  228. ├── CONTRIBUTING.md (contribution guidelines)
  229. ├── CHANGELOG.md (version history)
  230. │
  231. ├── vera-niw-skillset/ ← NIW skills (8 .skill files)
  232. │ ├── vera-niw-evaluate.skill
  233. │ ├── vera-niw-endeavor.skill
  234. │ ├── vera-niw-pillar.skill
  235. │ ├── vera-niw-recommendation.skill
  236. │ ├── vera-niw-assemble.skill
  237. │ ├── vera-niw-pl-review.skill
  238. │ ├── vera-niw-rfe-response.skill
  239. │ └── vera-niw-entrepreneur.skill
  240. │
  241. ├── vera-eb1-skillset/ ← EB-1 skills (11 .skill files)
  242. │ ├── vera-eb1-evaluate.skill
  243. │ ├── vera-eb1-authorship.skill
  244. │ ├── vera-eb1-original-contributions.skill
  245. │ ├── vera-eb1-judging.skill
  246. │ ├── vera-eb1-critical-role.skill
  247. │ ├── vera-eb1-published-material.skill
  248. │ ├── vera-eb1-recommendation.skill
  249. │ ├── vera-eb1-final-merits.skill
  250. │ ├── vera-eb1-assemble.skill
  251. │ ├── vera-eb1-pl-review.skill
  252. │ └── vera-eb1-rfe-response.skill
  253. │
  254. ├── vera-niw.plugin ← Bundled NIW plugin
  255. ├── vera-eb1.plugin ← Bundled EB-1 plugin
  256. │
  257. └── GoogleScholar/ ← Citation data tool
  258. ├── scholar.py
  259. ├── scholar_colab_demo.ipynb
  260. └── requirements.txt
  261. ```
  262. ---
  263. ## FAQ
  264. **Is this legal advice?**
  265. No. See [DISCLAIMER.md](DISCLAIMER.md). These tools provide informational guidance only. Always consult a qualified immigration attorney for your specific case.
  266. **Do I need to be technical?**
  267. No. Claude users can install plugins or upload `.skill` files. Codex users can extract the same `.skill` archives into their local skills directory.
  268. **Will this assess my approval chances?**
  269. 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.
  270. **How is this different from ChatGPT prompts for NIW?**
  271. 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.
  272. **Can I use this with Codex, GPT-4, or other models?**
  273. 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.
  274. **Can I contribute?**
  275. Yes. See [CONTRIBUTING.md](CONTRIBUTING.md).
  276. ---
  277. ## Contributing
  278. Contributions welcome. See [CONTRIBUTING.md](CONTRIBUTING.md) for details. Especially valuable:
  279. - **Bug reports** — Skill produced incorrect or misleading guidance? Open an issue.
  280. - **Test cases** — Real RFE patterns or edge cases for `evals/`. Anonymize all personal information.
  281. - **Reference materials** — New AAO decisions, policy updates, or adjudication trend data.
  282. - **Skill improvements** — Better rubrics, additional failure patterns, improved prompts.
  283. ---
  284. ## Disclaimer
  285. 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.
  286. ---
  287. ## License
  288. GPL-3.0 License. See [LICENSE](LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

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