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 agentmods add commands/mtvrkan/senior-dev-kit/design-checkgit clone --depth 1 https://github.com/mtvrkan/senior-dev-kitWrote 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/commands/mtvrkan/senior-dev-kit/design-check)<a href="https://agentmods.dev/commands/mtvrkan/senior-dev-kit/design-check"><img src="https://agentmods.dev/badge/commands/mtvrkan/senior-dev-kit/design-check.svg" alt="Measured on agentmods" 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 | $0.00017 | $0.01291 |
| Opus 5 | $0.00009 | $0.00646 |
| Sonnet 5 | $0.00003 | $0.00258 |
| Haiku 4.5 | $0.00002 | $0.00129 |
Grade A, and why
design-check 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 4d 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/design-check
Audit what was actually built against the design direction it claims: $ARGUMENTS
This is the counterpart to /seo-check. It does not review correctness — /code-review does that
— it answers one question: does this look like its own product, or like every other generated UI?
Read agent_docs/design-directions.md first; the direction table, the levers and the tells list
below are all defined there rather than restated here.
Step 1 — Find the claim. Read DESIGN-SPEC.md (project root) — the direction, the brief's
recorded constraints and exclusions, and the signature moment. No spec on a project that has UI is
itself the top finding: the design was never chosen, so it defaulted. Report it and derive the
de facto direction from the code so the rest of the audit still has something to measure against.
Step 2 — Direction adherence. For each axis in the spec — type, colour, geometry, depth,
density, motion, decoration, layout rhythm — check the code actually implements the recorded value.
The common failure is a spec naming one direction over a UI built from framework defaults: radius
still rounded-lg everywhere, one shadow on every surface, the body font never changed.
Step 3 — The tells. Scan for the generic-output signals in design-directions.md § THE TELLS.
Three or more means the UI will be read as machine-made whatever the tokens say. Report which ones
fired, with file:line.
Step 4 — Layout monotony. Count distinct section compositions across the page or flow. Three consecutive sections of "centred heading over an N-column card grid" is one idea repeated. Check the spec's layout rhythm is visible at all.
Step 5 — Depth coherence. Exactly one depth model should be in use (flat / soft shadow / hard offset / glow / glass). Shadow and glow and border and gradient on the same surface is the commonest way a chosen direction dissolves back into the default.
Step 6 — Consistency of the system. Count distinct values actually used for radius, shadow,
font size and section spacing. More values than the token scale defines means the scale is
decorative. Flag every hardcoded hex, text-[17px], p-[13px] and raw text-gray-*.
Step 7 — The invariants. Contrast against the real composited background (glass and saturated
colour fields fail here first), visible focus ring ≥3:1, prefers-reduced-motion honoured, target
size, the three mandatory states. A distinctive UI that fails these is not a win. This is a
sanity check on the design's own risk areas, not an accessibility audit — /a11y-check is that,
and a direction that scores well here can still be unusable by keyboard.
Step 8 — The signature. design-directions.md § THE SIGNATURE requires exactly one idea the
work is remembered for, executed in one place. Find it in the code: is the spec's signature moment
actually built, does it survive 360px and reduced-motion, and is the rest of the page quiet enough
for it to land? Absent is the ceiling finding — everything else can pass and the result is still
merely competent. More than one is the same finding from the other side: they cancel.
Step 9 — The brief. Every constraint and exclusion recorded in the spec, checked against the code: the brand palette that was fixed, the competitor to avoid, the density that was required. A UI that is distinctive but not what the user asked for has failed at the only bar that matters.
Output format:
DESIGN AUDIT
============
SPEC: [direction name from DESIGN-SPEC.md | NONE — never chosen]
DE FACTO: [what the code actually implements]
BRIEF: [constraints/exclusions honoured ✓ | ✗ [which one, file:line] | none recorded]
DRIFT (spec says X, code does Y):
✗ [axis] — spec: [value] · code: [value] — [file:line]
TELLS FIRED: [n]/11
✗ [tell] — [file:line]
MONOTONY:
⚠ [n] sections, [m] distinct compositions
⚠ depth models in use: [list] — should be exactly 1
SYSTEM:
⚠ radius values in use: [n] · shadow: [n] · font sizes: [n] · section spacing: [n]
✗ [hardcoded value] — [file:line]
INVARIANTS: contrast [✓/✗] · focus [✓/✗] · reduced-motion [✓/✗] · targets [✓/✗] · 3 states [✓/✗]
SIGNATURE: [the one idea, and where — | MISSING | [n] competing]
VERDICT: [Memorable / Distinctive / Competent but generic / Defaults wearing a direction's name]
TOP FIXES (highest visual return first):
1. [most impactful]
2. [second]
3. [third]
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.
- 4d ago First seen · 100 lines · 17 tokens per session scan A 8c479992a8c0
design-check is a command published in the GitHub repository mtvrkan/senior-dev-kit (5 stars, last pushed 4d ago), licensed MIT. It adds 17 tokens to every session and 1,291 once invoked, about $0.0001 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.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.