setuplisted

Use when the user runs /setup, asks to configure evolve-loop, onboard, pick per-phase models, or learn how the pipeline works. Detects which LLM CLIs/subscriptions are available, explains the pipeline concisely, proposes a per-phase model assignment the user can adjust, writes .evolve/llm_config.json, and validates it against the integrity floor. Runs once on first launch (the loop nudges) and is re-runnable anytime.
mickeyyaya/evolve-loop · ★ 1 · AI & Automation · score 77

Install: claude install-skill mickeyyaya/evolve-loop

# /setup > Interactive onboarding. The deterministic parts (detection, validation, marker) run in the Go binary (`evolve setup detect|validate|complete`); the judgment parts (recommend models, explain the pipeline) run HERE, in your session — zero extra API cost. You PROPOSE; `evolve setup validate` is the kernel CLAMP (envelope + allowed_clis are hard, cross-family is advisory). See [docs/architecture/setup-onboarding.md](../../docs/architecture/setup-onboarding.md) and the deterministic core in [go/internal/setup/setup.go](../../go/internal/setup/setup.go). ## When to use - The loop printed `[setup] First run …`, or the user typed `/setup`, or asked to configure models / learn the pipeline. - Re-running is always safe — it re-detects and rewrites `.evolve/llm_config.json`. ## Binary Call `evolve` if on PATH; otherwise `./go/bin/evolve` (or `$EVOLVE_GO_BIN`). All commands below are read-only except `complete` (stamps the marker) and your Write of `llm_config.json`. ## Procedure 1. **Detect.** Run `evolve setup detect --json` and parse it. The digest has `clis[]` (per family: `binary_present`, `auth_mode`, `subscription_type`, `capability_tier`, `verdict`, and **`tier_models`** {fast,balanced,deep} → that CLI's NATIVE model) and `phases[]` (per role: `current_cli`, `current_model`/`current_tier`, `source`, `envelope` {min,default,max}, `cross_family_with`, `allowed_clis`). 2. **Present the detection** as a compact table — one row per CLI family with binary/auth/tier/v