diff --git a/README.md b/README.md index b35687d8c..d38248e71 100644 --- a/README.md +++ b/README.md @@ -20,6 +20,18 @@ It's not a chatbot wearing hats. Each team member runs in its own context, reads > **Responsible AI stance** — Squad is built to amplify a human operator with GitHub Copilot, not to remove humans from the loop. Use it to delegate faster, review better, and keep governance close to the code. +### See it in Claude Code + +![Squad's team dashboard following a Claude Code session: the coordinator sends Dallas (Node.js engineer) to build a CLI and Parker (tester) to write tests at the same time, then Jordan (lead) reviews and Scribe records the decisions](docs/src/assets/images/squad-claude-dashboard.gif) + +A real run, sped up 12×: a fresh repository, `squad init --client claude`, then two prompts to `claude --agent squad`. "Cast a small team" ran `squad cast` for a lead, an engineer and a tester in 50 seconds. "Team, build the MVP" sent the engineer and the tester out in parallel, then the lead reviewed. The result was a commit with 22 passing `node --test` tests, about $2.60 of model time for both prompts. The page is `squad dashboard`, which reads Claude Code's own session files and needs no setup. + +```bash +squad init --client claude +claude --agent squad --mcp-config .mcp.json # "Cast a team for this repo", then give it work +squad dashboard --open # watch the team in your browser +``` + --- ## Quick Start diff --git a/docs/src/assets/images/squad-claude-dashboard.gif b/docs/src/assets/images/squad-claude-dashboard.gif new file mode 100644 index 000000000..39ff69c12 Binary files /dev/null and b/docs/src/assets/images/squad-claude-dashboard.gif differ