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 instructions/ssheleg/super-ux/claude-mdgit clone --depth 1 https://github.com/ssheleg/super-uxWrote 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/instructions/ssheleg/super-ux/claude-md)<a href="https://agentmods.dev/instructions/ssheleg/super-ux/claude-md"><img src="https://agentmods.dev/badge/instructions/ssheleg/super-ux/claude-md.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 5.1 | $0.02124 | $0.02124 |
| Opus 5 | $0.01062 | $0.01062 |
| Sonnet 5 | $0.00425 | $0.00425 |
| Haiku 4.5 | $0.00212 | $0.00212 |
Grade A, and why
super-ux CLAUDE.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 5d 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 — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — super-ux
Guidance for agents working on super-ux. For using it, see README.md.
super-ux is the plugin that demands a design chain before UI code. It has a
user-facing interface of its own — the installer CLI and its interactive
menu in bin/super-ux.js — so it is bound by its own rules. It went three
months without them; the 2026-08-10 audit found four defects in that CLI
inside an hour of writing the scenarios it should have had.
Where the artifacts are
| Path | What |
|---|---|
docs/ux/ |
This project's own chain — vision, foundation, flows, screens, scenarios |
docs/brand/ |
This project's own verbal identity, and the surfaces it governs |
docs/evidence/ |
Pipeline records: briefs, specs, plans, retro.md, backlog.md, verification.md |
plugins/super-ux/skills/references/ |
Source of truth for every contract. After editing, run python3 test/sync_references.py |
templates/ |
Seeds seen by target projects — including both hard rules, whose only source these are. Source of truth: shipped copies live in plugins/super-ux/templates/ and inside each skill that seeds from one; after editing, run python3 test/sync_references.py |
docs/AGENT_SYNC.md |
How coordination is wired here, and what it does NOT guarantee. Generated from .claude/agent-sync.json — read it before editing a guarded file, and regenerate it in the same change that alters the config |
Gates
Run each alone and read its own exit code. Never through a pipe: tail
exits 0 whatever it reads, so validate.py | tail -2 && git commit commits
on a red validator. That happened once, in this family, on a tag.
python3 test/validate.py # repo consistency
python3 test/brand_lint_test.py # one fixture per brand-lint code
python3 docs/ux/lint.py # this project's own chain
python3 docs/brand/lint.py # this project's own copy
python3 test/release_preflight.py # before any tag
Editing anything under plugins/super-ux/skills/references/ — or under
templates/ — requires python3 test/sync_references.py in the same
change: the per-skill copies are what non-Claude agents actually receive,
and the plugin's templates/ mirror is what installed channels seed from.
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.
- 5d ago First seen · 153 lines · 2,124 tokens per session scan A df276d304acb
super-ux CLAUDE.md is an instructions file published in the GitHub repository ssheleg/super-ux (3 stars, last pushed yesterday), licensed MIT. It adds 2,124 tokens to every session, about $0.0106 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 instructions, from other repositories
figma-console-mcp CLAUDE.md
Claude Code instructions for southleft/figma-console-mcp, covering figma console mcp, build & test, release process, known issues and architecture.
FireCrasher CLAUDE.md
Instructions for osama-raddad/FireCrasher: The guidance for AI assistants in this repository lives in AGENTS.md, the vendor-neutral instructions file. It is the single source of truth — read it instead of this file.
get-shit-pretty CLAUDE.md
Claude Code instructions for jubscodes/get-shit-pretty, covering gsp — get shit pretty, must-always / must-never, git workflow, this repository and architecture.
phonebook CLAUDE.md
Claude Code instructions for stag-build/phonebook, covering @stag-build/phonebook and model/agent delegation (token budget).
funnel-architect-plugin CLAUDE.md
Instructions for ominou5/funnel-architect-plugin, covering funnel architect, how to delegate, workflow, funnel hacking workflow and skills (in skills/).
design-sandbox CLAUDE.md
Instructions for memi-design/design-sandbox, a project described as: memi-ready Next.js, Tailwind, shadcn, MCP, and Agent Skills sandbox for interface understanding and design-to-code exploration.