HEAD_SCREEN_ONCE=1 state/bin/head-screen/launch. .md file to compare - side-by-side diff against snappy-face
snappy-face
description: "Triggers on prompt mention of 'snappy-face'."
What it does for you
Opens your assistant's live console on your screen.
What it produces
A recent result, so you can see the kind of work it returns.
loading…
How to get it
These run inside the Snappy workspace. Want this working in your business? I set skills like this up with you, in one focused week.
For developers how this skill is built, graded, and how it runs
at a glance- the short version
what's inside - the parts that make up a skill 2/4 present
A skill is just a few plain-text files. Only the main one is required. The rest are optional, added as the work needs them. This is what the skill is made of; how it runs is just below.
state/skills/snappy-face/SKILL.md
present
state/lib/snappy-face.ts
not present
state/bin/snappy-face/
not present
state/skills/snappy-face/AGENTS.md
present
how it's graded - what counts as a good run 3 criteria · 2 deterministic · 1 judge
Each row is one thing a good run has to get right. deterministic means a quick check decides, pass or fail. judge means the AI reads the result and rates it. Grading each piece on its own (instead of one overall score) shows exactly where a run fell short, so the fix is obvious.
how it runs - the shared frame every skill uses 3/5 present
Every skill runs the same way. One part does the work, a separate part checks it, and a short loader hands the AI exactly what it needs for the job. Anything this skill doesn't use shows a one-line note saying why, on purpose, not by accident.
state/log/evals.ndjson - Idempotent: if server is up (/healthz 200) and menubar is running (pgrep snappy-menubar), just confirm the window is visible and stop.
- Always pass HEAD_SCREEN_ONCE=1 so launch.sh returns after the window opens.
- Do not start the server in the foreground — it blocks the turn.
- For window-targeted screen capture of non-cockpit apps from headless / SSH / claude --dangerously-skip-permissions chains, direct osascript / screencapture / cliclick silently fail TCC because the responsible client at runtime resolves to claude.exe (no Screen Recording grant). Use peekaboo from its canonical /opt/homebrew/Cellar/peekaboo/3.0.0-beta3/bin/peekaboo install (TCC granted to the canonical path). Pattern:
what it has learned - fixes written back in over time sample
When a run hits something this skill didn't handle, the fix gets written back into the skill so it doesn't happen again. FIXED means it was corrected on the spot. LOGGED means it's queued for a bigger rewrite. Either way, the skill gets a little better and never makes the same mistake twice.
- Loading feedback rows…
how the work flows- step by step
SKILL.md- the skill, written out in plain English
snappy-face
One job: get the operator console on-screen with one command. Starts the head-screen server on port 3147, spawns the snappy-menubar status-bar icon, and opens the console in a Chrome --app window. Idempotent - each step is skipped if its process is already running.
Usage
/snappy-face
Steps
- Run:
HEAD_SCREEN_ONCE=1 state/bin/head-screen/launch.sh
- The script will:
- Start
state/bin/head-screen/server.tson127.0.0.1:3147if/healthzdoesn't respond. - Spawn
state/bin/head-screen/snappy-menubarifpgrepdoesn't find it. - Open the console in Chrome
--appmode (menubar handles this on first launch; otherwiselaunch.shopens it directly).
- Confirm the ⚡ icon is visible in the Mac status bar and the console
window is on-screen.
Why
Three things need to coexist for the console to work: the HTTP server, the menubar health indicator, and the browser window. Running them independently means forgetting one. This wraps all three.
Eval
Shape-gate:
curl -fsS http://127.0.0.1:3147/healthzreturns 200.pgrep -f snappy-menubarreturns a PID.
Score = 1 if both pass; else 0.
Gotchas
- Requires the menubar binary to have been built once:
bash state/bin/head-screen/build-menubar.sh. The repo ships the binary at state/bin/head-screen/snappy-menubar, so this is only needed if it's missing or stale.
HEAD_SCREEN_ONCE=1is important - without it,launch.shstays
foregrounded waiting on the server; with it, the launcher exits once the window is up and the server keeps running in the background.
- To stop everything:
pkill -f "tsx.*head-screen/server"and
pkill -f snappy-menubar.
Rubric
criteria:
- name: server_healthz_responds
kind: deterministic
check: "The command 'curl -fsS http://127.0.0.1:3147/healthz' returns a 200 OK status code."
- name: menubar_process_running
kind: deterministic
check: "The command 'pgrep -f snappy-menubar' returns a Process ID (PID)."
- name: console_window_visible
kind: judge
check: "The operator console window is visibly displayed on the screen as a Chrome --app window."AGENTS.md- what the AI loads when this skill comes up
snappy-face - loader
Per-turn rules for the snappy-face skill. Full reference: state/skills/snappy-face/SKILL.md. Do not skip these.
Critical Rules
- Idempotent: if server is up (
/healthz200) and menubar is running (pgrep snappy-menubar), just confirm the window is visible and stop. - Always pass
HEAD_SCREEN_ONCE=1solaunch.shreturns after the window opens. - Do not start the server in the foreground - it blocks the turn.
- For window-targeted screen capture of non-cockpit apps from headless / SSH /
claude --dangerously-skip-permissionschains, directosascript/screencapture/cliclicksilently fail TCC because the responsible client at runtime resolves toclaude.exe(no Screen Recording grant). Usepeekaboofrom its canonical/opt/homebrew/Cellar/peekaboo/3.0.0-beta3/bin/peekabooinstall (TCC granted to the canonical path). Pattern:
# 1. find window id by app name + bounds
/opt/homebrew/bin/peekaboo list windows --app "<App>" --json
# 2. capture by window id
/opt/homebrew/bin/peekaboo image --app "<App>" --window-id <id> --retina --path /tmp/cap.png
No Peekaboo.app/Claude.app MCP bridge socket exists at ~/Library/Application Support/{Peekaboo,Claude,clawdis}/bridge.sock; CLI is the only reachable surface. (writeback 2026-04-29T01:18:10Z)
Commands
| ui dashboard | state/skills/snappy-face/resources/ui.openui | |invoke: HEAD_SCREEN_ONCE=1 state/bin/head-screen/launch.sh |capture-window (TCC-safe): /opt/homebrew/bin/peekaboo image --app "<App>" --window-id <id> --retina --path <out> |find-window: /opt/homebrew/bin/peekaboo list windows --app "<App>" --json |eval log: state/log/evals.ndjson (skill: "snappy-face")
Self-Test
An agent reading this should correctly:
- [ ] Know the backing script is
state/bin/head-screen/launch.sh - [ ] Know what to write to
state/log/evals.ndjsonafter invoking (shape-gate on/healthz+pgrep snappy-menubar) - [ ] Know the eval mode is
auto-shape
Self-report
If this loader fell short, append a line:
echo "[$(date -u +%FT%TZ)] snappy-face: <what was missing>" >> state/log/loader-feedback.log
<!-- kernel-ok: historical attribution for the PID-loader pattern. -->
Self-correcting loader (PID feedback)
This is the rule that makes loaders learn. Ported from snappy-kernel, where it demonstrably worked: the snappy-dom-cartographer skill mapped Skool admin UI once, every later ship-agent inherited the full map for free.
Before you finish, do two things:
1. Fix gaps (P - proportional). MANDATORY. If this loader didn't cover your case - if you had to read another file, run an undocumented command, work around a wrong selector, or discover a quirk - you MUST attempt an Edit to this AGENTS.md before you log.
- Default action: edit this file. Surgical. Add the missing line, table
row, or code example where it belongs. No restructuring.
- Only
LOGGEDis allowed when: the fix needs >10 lines, spans multiple
files, or requires a structural rewrite. In that case the state/regen/drain.sh queue picks it up asynchronously.
- "I didn't have time" / "it's minor" / "the next agent will figure it out"
are NOT valid reasons. Minor gaps compound - that's the whole point of this loop.
- Concrete example of a valid P-fix: discover
agent-browser screenshot
takes path positionally not as --output <path>. One line in a table. Edit it in. Do not just log it.
- The goal: the next agent never has to leave the loader.
2. Log the result.
echo "[$(date -u +%FT%TZ)] snappy-face: <what was missing or fixed> [FIXED|LOGGED]" >> state/log/loader-feedback.log
FIXED= you patched this loader inline (P-fix).LOGGED= too large for inline; the PostToolUse enqueue + Stop-hook drain
will rewrite the loader from scratch on next session-end.
Do not skip this. Every agent run must leave the system better than it found it. The loader is the setpoint; you are the sensor; the gap is the error signal; closing the gap is the correction.
OpenUI Resource
- Skill-owned OpenUI Lang resource:
state/skills/snappy-face/resources/ui.openui. Read it before rendering or editing this skill's generated component surface. - Treat this resource as a first-class artifact of the skill, not a generic chat response. Improve it when the skill's user-facing output needs to become richer.
- System resources compose OpenUI primitives and inherit SnappyChat tokens. Use
ui_contract: brandedin SKILL.md only for deliberate platform or client visuals.
api.ts- the code it can call
⚠ no api.ts - this skill has no typed action surface
scripts- helper scripts it can run
prose-only skill - 2 inline code blocks live in SKILL.md above (no state/bin/ sidecar yet).
how we check it- the checks, plus the last 4 runs
| timestamp | verb | score | primary_issue | artifact |
|---|---|---|---|---|
| 2026-04-25 04:11Z | - | 1.00 | - | - |
| 2026-04-25 04:11Z | - | 1.00 | - | - |
| 2026-04-25 04:11Z | - | 1.00 | - | - |
| 2026-04-25 04:11Z | - | 1.00 | - | - |