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.
npx skills add OthmanAdi/MDDesign --skill design-harvestgit clone --depth 1 https://github.com/OthmanAdi/MDDesignWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/othmanadi/mddesign/design-harvest)<a href="https://agentmods.dev/skills/othmanadi/mddesign/design-harvest"><img src="https://agentmods.dev/badge/skills/othmanadi/mddesign/design-harvest/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/othmanadi/mddesign/design-harvest"><img src="https://agentmods.dev/badge/skills/othmanadi/mddesign/design-harvest.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 5.1 | $0.00083 | $0.01995 |
| Opus 5 | $0.00042 | $0.00997 |
| Sonnet 5 | $0.00017 | $0.00399 |
| Haiku 4.5 | $0.00008 | $0.00199 |
Grade A, and why
design-harvest 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 9d 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.
How it starts
The opening of the file, as written. The whole thing — 173 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Design Harvest
The on-ramp. Three modes.
Mode 1: harvest (/mddesign:harvest)
Goal: produce a structurally valid DESIGN.md at project root from the existing codebase.
Step 1: Pre-flight
- Glob for
DESIGN.mdat project root. - If it exists and the user did NOT pass
--overwrite, refuse: print "DESIGN.md already exists. Use/mddesign:critiqueinstead, or pass--overwriteto replace." - Confirm
npxis available (bash -c "command -v npx").
Step 2: Dispatch the design-archeologist subagent
Call the Agent tool with subagent_type: design-archeologist and a prompt containing this PhaseSpec:
phase_id: harvest-<short-uuid>
parent_plan_ref: ./task_plan.md
goal: Extract de-facto design tokens from this codebase and return a draft DESIGN.md body following the Google Labs spec.
done_when:
- draft_design_md is non-empty
- draft_design_md contains YAML frontmatter and 8 ordered markdown sections
- draft_design_md includes a section #9 "Agent Prompt Guide"
inputs:
files: [] # the subagent globs the project itself
tools_allowed: ["Read", "Glob", "Grep", "Bash"]
tools_denied: ["Write", "Edit", "WebFetch"]
budget:
max_tool_calls: 200
max_wall_seconds: 600
hitl_checkpoints:
- before_returning_draft
return_contract:
format: markdown+frontmatter
fields: [summary, draft_design_md, open_questions]
The subagent has its own dedicated SKILL/agent definition at agents/design-archeologist.md.
Step 3: Receive the draft, write to DESIGN.md
- Take the
draft_design_mdfield from the subagent's return. Writeit to<project_root>/DESIGN.md. This write goes through the PreToolUse HITL gate (named checkpoint:before_design_md_write).
Step 4: Lint and iterate
Run:
npx --yes @google/design.md lint DESIGN.md
If exit code is 0, you are done.
If non-zero, capture the linter output. Re-dispatch design-archeologist with an additional input field lint_errors: <output> and instruction "fix every reported error and return a new draft." Loop up to 3 times. After 3 failed iterations, surface the final lint errors to the user and stop.
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.
- 9d ago First seen · 173 lines · 83 tokens per session scan A b6cad802bfb0
design-harvest is a skill published in the GitHub repository OthmanAdi/MDDesign (13 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 83 tokens to every session and 1,995 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
a11y-validate
Accessibility validator: WCAG 2.1 AA, EN 301 549, EAA. Triggers: a11y, accessibility, WCAG, EAA, ARIA, contrast, keyboard, screen reader.
design-engineering
UI craftsmanship: animation rules, easing, micro-interactions, state polish. Triggers: animation, transition, ease-out, motion, micro-interaction, hover, loading state, UI polish.
johnny-suede-design
Suede Labs AI full-stack surface builder that runs design, copy, and visual QA as one pass: landing pages, brand surfaces, product UI, dashboards, campaigns, launch pages, and reference-to-target restyles (suedify). Use when a build needs layout and words together, when a redesign or launch surface has to ship end to…
suede-site-alchemy
Suede-owned conversion-path discipline for pages. Turns a page into a conversion path: hero, friction, proof, CTA, pricing, A/B ideas, quick wins, and mobile CRO. Use when a live or drafted page needs conversion-rate work on its structure and persuasion path. NOT FOR: a scored SEO/AI-visibility audit (use…
suede-design
Suede Labs AI design skill for making an interface feel intentional instead of templated: design tokens, color strategy, OKLCH ramps, type scale, fluid type, visual hierarchy, dark mode, spacing, component laws, motion, and source-to-implementation visual QA. Use when asked to design, restyle, polish, or audit a…
frontend-design-review
Design, review, and visually QA web and app interfaces against explicit design laws so shipped screens look intentional instead of generic AI output.