What is vibe-hardware AGENTS.md?

A repository instruction guide for AI coding agents working on Vibe Hardware, a project for designing and building hardware products. It maps the repository and directs agents to the detailed instructions for each hardware area.

Use it when modifying the Vibe Hardware repository, especially when deciding which skill covers firmware, circuit-board design, CAD, user interfaces, industrial design, or voice-guided physical work.

It gives an agent the project structure, required checks, and rules for choosing the right instructions before making changes. This reduces mistakes caused by applying the wrong process to firmware, circuit boards, computer-aided design, or industrial design.

vibe-hardware AGENTS.md — Instructions file by luckiday

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add instructions/luckiday/vibe-hardware/agents-md

Clone the repo

git clone --depth 1 https://github.com/luckiday/vibe-hardware

Made for: Codex, OpenCode.

Per session 1,668 This file is loaded in full into every session.

When invoked 1,668 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.01668	$0.01668
Opus 5	$0.00834	$0.00834
Sonnet 5	$0.00334	$0.00334
Haiku 4.5	$0.00167	$0.00167

Measured 2d ago against content hash 7479370f3746, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

vibe-hardware AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 106 lines

How it starts

The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Guidance for AI agents working in this repo. Keep this short — the detailed method and gotchas for any task live in the relevant skill's SKILL.md + references/. Read those before acting; this file is the map and the rules.

Repo map

skills/          the agent skills (one per dir):
  vibe-plm              the integration layer: product manifest + interface contracts
  vibe-firmware · vibe-pcb · vibe-cad   the three domains (each a self-contained loop)
  vibe-uiflow           firmware's cheaper operating point: UIFlow2/MicroPython UI devices
  vibe-industrial-design  appearance before structure: fixed constraints → AI look image
                        (received from the owner, or generated from the current render) →
                        measured params → three.js scene → contact sheet → Blender finals
                        → versioned ID report
  vibe-voice            the human channel: speak the hands-on steps (bring-up, probing,
                        test-fitting) instead of writing them where they won't be read
  <skill>/SKILL.md      frontmatter (name + "use when…") + the method
  <skill>/references/   deep lessons / gotchas, loaded on demand
  <skill>/scripts/      portable bash/python tools (plm_check, pcb_check, pcb_view, cad_viewer, …)
examples/        worked end-to-end builds (pager-buddy, …); product.yaml + the three dirs
docs/            getting-started and friends

The three domains are coupled only through data — the contract files named in a product's product.yaml (vibe-plm owns it), never by one skill calling another's code.

Repo rules

Keep root guidance short. Put domain detail in the skill it belongs to, not here.
Skills are self-contained. A skill's code (its scripts/) must not import or call another skill's code. Prose "hand off to the sibling skill" pointers are fine; a runtime dependency between skills is not. (vibe-pcb's viewer renders in-page; it does not shell out to vibe-cad.)
Generate by script, review in the browser. Sources of truth are small generator/parametric files (KiCad via gen_*.py; build123d .py) — never hand-edited GUI binaries. Review interactively on localhost, never by booting a heavyweight GUI. If a change needs the GUI to use, it isn't done.
Keep it general; no proprietary designs. Skills teach a reusable method, illustrated with generic parts. Worked product designs go in examples/ (only what you choose to publish) — never bake a private design into a skill.
Living docs. When a real build teaches a new gotcha/fix/better-practice, fold it back into the skill and commit it with the work. Compounding the experience is the point.
Global by default. Non-English text is fine where it belongs — language fixtures in code, quoted strings, log excerpts, translated docs. Keep the shared docs readable in English, and put a translation in a *.<lang>.md beside its original rather than in place of it.

Read the full file on GitHub · 106 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2d ago First seen · 106 lines · 1,668 tokens per session scan A 7479370f3746

Subscribe to this mod's changes

vibe-hardware AGENTS.md is an instructions file published in the GitHub repository luckiday/vibe-hardware (5 stars, last pushed 10d ago), licensed MIT. It adds 1,668 tokens to every session, about $0.0083 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Hardware and IoT

vibe-hardware AGENTS.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

AGENTS.md

Repo map

Repo rules

What this file has done since we first saw it

Other instructions, from other repositories

KernelWiki CLAUDE.md

tilelang-ascend AGENTS.md

openbridge-webcomponents automation-components.instructions.md

Vanchor AGENTS.md

brilliant_sdk AGENTS.md

ROS AGENTS.md