|
| 1 | +# PyAutoBrain command surface — shared reference |
| 2 | + |
| 3 | +The short verb commands (`/feature`, `/build`, `/health`, `/bug`, `/refactor`, |
| 4 | +`/docs`, `/research`, `/route`, `/brain`) are a thin, human-friendly **veneer** |
| 5 | +over the PyAutoBrain router (`bin/pyauto-brain`). This file is the shared context |
| 6 | +every command file points at, so each command body stays a few lines long. |
| 7 | + |
| 8 | +> **Users speak in short commands; PyAutoBrain performs the routing.** |
| 9 | +
|
| 10 | +## The Brain is implicit |
| 11 | + |
| 12 | +You don't tell a brain to "activate the visual cortex" — you *look*, and it |
| 13 | +routes. Same here: a user types a short verb (or plain natural language) and the |
| 14 | +Brain routes it to the right specialist agent. **Normal usage never says |
| 15 | +"PyAutoBrain".** The only explicit-Brain surface is `/brain`, the debug door. |
| 16 | + |
| 17 | +## Never bypass the Brain |
| 18 | + |
| 19 | +Every command routes **through** PyAutoBrain — either its CLI (`bin/pyauto-brain |
| 20 | +<agent>`) or the `start_dev` workflow entry point, which itself routes reasoning |
| 21 | +through the Brain Feature Agent. A command is a *shortcut into* the Brain, never a |
| 22 | +*replacement for* it. No command file re-implements classification, planning, the |
| 23 | +readiness gate, or execution — those belong to the organs. |
| 24 | + |
| 25 | +## The three command tiers |
| 26 | + |
| 27 | +**1. Real conductors** — route straight to an existing agent in |
| 28 | +`agents/conductors/` (`AGENTS.md` is authoritative): |
| 29 | + |
| 30 | +| Command | Agent | Chain | |
| 31 | +|---------|-------|-------| |
| 32 | +| `/feature` | Feature Agent | `bin/pyauto-brain feature` → `start_dev` → `ship_*` | |
| 33 | +| `/build` | Build Agent | `bin/pyauto-brain build` → vitals faculty → Heart → PyAutoBuild | |
| 34 | +| `/health` | Health Agent | `bin/pyauto-brain health` loop → vitals faculty → Heart → GREEN | |
| 35 | + |
| 36 | +**2. Work-type entries** — no dedicated conductor exists **yet**, so these route |
| 37 | +through the Brain dev-flow with their PyAutoMind work-type fixed. Still through |
| 38 | +the Brain (via `start_dev` → Feature Agent), so nothing is bypassed: |
| 39 | + |
| 40 | +| Command | PyAutoMind work-type | Promotion follow-up | |
| 41 | +|---------|----------------------|---------------------| |
| 42 | +| `/bug` | `bug/` | dedicated Bug conductor | |
| 43 | +| `/refactor` | `refactor/` | dedicated Refactor conductor | |
| 44 | +| `/docs` | `docs/` | dedicated Documentation conductor | |
| 45 | +| `/research` | `research/` | dedicated Research conductor | |
| 46 | + |
| 47 | +These are honest interim doors — they do **not** pretend an agent exists that |
| 48 | +doesn't. The taxonomy they tag is `PyAutoMind/ROUTING.md`. |
| 49 | + |
| 50 | +**3. Router + debug door:** |
| 51 | + |
| 52 | +- **`/route <free text>`** — the natural-language door (the "look at this" path): |
| 53 | + infer the work-type from the request and dispatch to the matching command |
| 54 | + above. This is the primary interface; the verbs are typed shortcuts into it. |
| 55 | +- **`/brain <agent> [args]`** — explicit, un-veneered passthrough to |
| 56 | + `bin/pyauto-brain` for debugging. Free-text `/brain` defers to `/route`. |
| 57 | + |
| 58 | +## How these are installed |
| 59 | + |
| 60 | +Each command is a directory `skills/<verb>/` containing only `<verb>.md` (no |
| 61 | +`SKILL.md`). `bin/install.sh` turns that into a flat `~/.claude/commands/<verb>.md` |
| 62 | +symlink — a typed slash command, not an auto-triggered skill. This file |
| 63 | +(`COMMANDS.md`) sits at the `skills/` root, so the installer's directory scan |
| 64 | +skips it: it is reference-only. Keep command bodies short and this file the single |
| 65 | +place architecture prose lives (guarded by `bin/check_skill_line_counts.sh`). |
0 commit comments