AI Dataset Generator
SQLite-backed pipeline that generates, verifies, dedupes, and exports SFT/DPO training datasets.
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 15, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 41e696b
Ran the real SQLite pipeline end to end (import 5 drafted Python-comprehension SFT records, adversarial LLM-judge review, dedup, export) and it produced a working flat_train.jsonl plus a genuinely useful DATA_CARD.md with judge-score and difficulty distributions - but the exact two-step verify command sequence printed in SKILL.md's own numbered workflow silently processes 0 records the second time because step 6 already moved records out of the raw_generated status step 7 filters on.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 9/10
- Docs & honesty 3/5
What AI Dataset Generator does
Runs a deterministic local pipeline (generate, heuristic + adversarial LLM-judge verify, shingle dedup, schema-driven export with a data card) for building SFT or DPO fine-tuning datasets from topics, URLs, or existing JSONL/CSV files. Triggers on requests to generate, clean, verify, or export training datasets for Codex, Antigravity, or Claude Code.
How to install AI Dataset Generator
git clone https://github.com/Bhanunamikaze/AI-Dataset-Generator
cd AI-Dataset-Generator
mkdir -p ~/.claude/skills
cp -r . ~/.claude/skills/dataset-generator
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger AI Dataset Generator
-
/dataset-generatorSQLite-backed pipeline that generates, verifies, dedupes, and exports SFT/DPO training datasets.
It also activates on plain-language prompts like these:
-
Generate a small SFT training dataset of Python Q&A pairs for fine-tuning -
Dedupe and verify this JSONL dataset before I export it for DPO training -
Build a fine-tuning dataset from these URLs with an adversarial judge pass
Frequently asked questions
- Is the AI Dataset Generator skill free?
- Yes. The skill itself is free from Bhanunamikaze/AI-Dataset-Generator. SkillProof publishes the install command and an independent test verdict at no cost.
- Does AI Dataset Generator work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 15, 2026. Verdict: Tested · Works. Ran the real SQLite pipeline end to end (import 5 drafted Python-comprehension SFT records, adversarial LLM-judge review, dedup, export) and it produced a working flat_train.jsonl plus a genuinely useful DATA_CARD.md with judge-score and difficulty distributions - but the exact two-step verify command sequence printed in SKILL.md's own numbered workflow silently processes 0 records the second time because step 6 already moved records out of the raw_generated status step 7 filters on.
- What is the AI Dataset Generator SkillProof Score?
- 8.8/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 9/10, docs & honesty 3/5.
- How do I install AI Dataset Generator?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use AI Dataset Generator with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.