Install & First Run

Prerequisites

• Node.js >= 18 — check with node -v
• A coding-agent CLI on your PATH — Claude Code or Codex:
  • Claude Code: npm install -g @anthropic-ai/claude-code (claude.com/claude-code)
  • Codex: npm install -g @openai/codex (github.com/openai/codex)

You need one of those on your PATH before running rolester start.

Public install

npm install -g rolester
rolester start claude        # or: rolester start codex

The start command does the whole arc in one shot:

1. Scaffolds candidate/ and workspace/ directories (idempotent — safe to re-run).
2. Installs the skills so your agent sees /apply-job, /evaluate-job, etc.
3. Seeds workspace/tracker.json from the demo template if not yet present.
4. Boots the live dashboard at http://localhost:7777 with hot reload.
5. Launches your agent with a starter message.

Source checkout

Developing from a source checkout uses the same command shape after linking the local binary:

git clone https://github.com/CodesWhat/rolester
cd rolester
npm install
npm link
rolester start claude        # or: rolester start codex

Useful flags

Manual wiring

If you prefer to open your agent yourself:

rolester doctor     # verify environment + scaffold state

Then open Claude Code or Codex in the repo root and say:

familiarize yourself and let's get started

The agent reads AGENTS.md, verifies the skills shim, and runs ingest-profile conversationally if your candidate profile is not yet set up.

Custom data location

By default workspace/ and candidate/ live inside the cloned repo. Set ROLESTER_HOME to put them somewhere else:

export ROLESTER_HOME=~/rolester-data
rolester start claude

Everything under ROLESTER_HOME is gitignored and never touches the repo tree.

Health check

rolester doctor

Reports environment health (Node version, agent CLI on PATH), skills discoverability, workspace scaffold state, setup completeness, and any config schema errors. Run this any time something feels off.