# PalatineAI

> The evaluation layer for new opportunities. PalatineAI researches, ranks, and tracks Colosseum hackathon projects (2,857 researched, 1,680 scored across 12 signals), and publishes the top 120 as an agent-readable dataset with an installable skill. All artifacts are static files over HTTPS — no auth, no rate limits.

## Data

- [projects.json](https://palatineai.site/projects.json): The dataset — top 120 scored projects with 12 signals each, plus database stats and the `stats.updated` version stamp. ~140 KB JSON.
- [skill.md](https://palatineai.site/skill.md): Agent skill (markdown + YAML frontmatter, v0.2.0) — teaches an agent the dataset schema, ready-made queries, and the evaluation workflow. Install for Claude Code: `mkdir -p ~/.claude/skills/palatineai && curl -fsSL https://palatineai.site/skill.md -o ~/.claude/skills/palatineai/SKILL.md`

## Site

- [Dashboard](https://palatineai.site/): Human-facing live rankings, filters, and the scoring model.

## Notes

- Each project: `project`, `description`, `category`, `country`, `team`, `website`, `x`, `scores` (12 signals, 0–10 each), `total` (max 120), `rank`, `tokenStatus`.
- Inverted signals: `scores.noToken = 10` means no token exists yet (opportunity); `scores.funding = 10` means not yet heavily funded.
- Not financial advice. Research tooling only.