← ClaudeAtlas

canary-lablisted

Use when the user asks Codex to run, verify, debug, heal, create, or export Canary Lab features through Canary Lab MCP tools. Guides Codex through external run repair plus author workflows with create_feature, env capture, external draft apply, and evaluation export.
ferterahadi/canary-lab · ★ 3 · AI & Automation · score 76
Install: claude install-skill ferterahadi/canary-lab
# Canary Lab Use Canary Lab MCP tools for Canary Lab runs instead of ad hoc shell commands when available. ## Workspace Bootstrap Before calling Canary Lab MCP tools, make sure the workspace and UI server are available. 1. Read the user-level registry at `~/.canary-lab/workspaces.json`. On Windows, resolve it from the user's home directory, for example `%USERPROFILE%\.canary-lab\workspaces.json`. 2. If the registry has exactly one workspace, use that workspace. If it has multiple workspaces, list their `name` and `path` values and ask which one to use. 3. If the registry is missing or empty, ask the user to run `npx canary-lab setup` from the Canary Lab workspace. 4. Check `http://127.0.0.1:7421/mcp/health`. 5. If the health check succeeds, confirm `projectRoot` matches the selected workspace. If it points at a different workspace, ask the user whether to stop the existing Canary Lab server before continuing. 6. If the health check fails, start `npx canary-lab ui` from the selected workspace in a visible long-running terminal when the host supports that. Do not add `--port`; Canary Lab uses port `7421` so MCP clients can connect consistently. 7. Do not reflexively call `list_features` or `list_runs` after health. For random or new feature creation, call `create_feature` directly with a unique feature name. Use `list_features` only when you need to discover or choose an existing feature, and use `list_runs` only for run, heal, verification, or export workflows. ## Externa