# Upstream packs this skill draws on

The interview, necessity test, scoring, signal and pivot references are vendored in this
folder, rewritten with attribution, so the skill needs no upstream pack to run. Two links
still point outside the repo, both pinned to a commit.

## Superteam ideas dataset

[superteam-ideas.json](https://github.com/sendaifun/solana-new/blob/e81c261645035c0e902eaaa518ff58722d188bb7/skills/data/ideas/superteam-ideas.json):
240 Solana-native idea entries as inert JSON, read over https. To grep it locally, download
that one file; nothing else from its repository is needed.

## colosseum-copilot (ColosseumOrg/colosseum-copilot)

Queries 5,400+ past Colosseum submissions for cluster density, winner patterns and gaps.
Backed by a CLI, so it needs a one-time sign-in per machine regardless of install route:

```
npx @colosseum-org/copilot-connect login
```

Node 20+ required.

- Upstream: https://github.com/ColosseumOrg/colosseum-copilot
- Pinned commit for the link in `SKILL.md`: `0453ffe26e8d245152619fcc949689a6adaab1c6`
- With [solanabr/ai-kit](https://github.com/solanabr/ai-kit) installed: `bash .claude/bin/skills.sh add colosseum` (installed by default there)