Menu
Bootstrap new repo experiments with the red conventions for folder layout, runtime config, testing, and compose-based e2e.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Cut a production release for red and verify it is healthy in production. Use when the user wants to tag and publish a release, monitor the GitHub Release workflow, confirm `red.computer` is healthy after deploy, or reason about rollback for a bad production release.
Full lifecycle for repo daemons — create, seed memory, run, and refine via interactive resume. Use when adding a new `*.daemon.md` audit agent, tightening an existing one, seeding a daemon's initial memory with tracked facts from authority files, or verifying a new daemon reasons correctly before CI rotation.
Debug a red preview deployment on the shared Hetzner box. Use when a preview URL, deploy, remote service, or per-PR stack needs investigation over SSH, or when you need to find where preview services, logs, caddy routes, env files, and compose state live on the host.
Use when debugging red request flow, verifying wide-events ingestion, inspecting MinIO raw or rollup data, checking request_id propagation, or quickly tracing why a request did or did not materialize into a canonical rollup.
Based on SOC occupation classification
| name | experiment-bootstrap |
| description | Bootstrap new repo experiments with the red conventions for folder layout, runtime config, testing, and compose-based e2e. |
Use this skill when starting a new experiment under experiments/, or when
standardizing an existing one onto the repo's experiment conventions.
experiments/<name>/.src/, compose/, README.md, Dockerfile, docker-compose.yml, package.json, tsconfig.json, and a local bun.lock.src/service/ for runtime services and adapterssrc/store/ for persistence abstractionssrc/util/ for config, errors, and shared typessrc/test/ for tests and fixturesjust for commands.Read the experiment checklist before scaffolding or refactoring an experiment.