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/mouse-lin/finesse-skill/copilot-instructionsgit clone --depth 1 https://github.com/mouse-lin/finesse-skillWrote 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/mouse-lin/finesse-skill/copilot-instructions)<a href="https://agentmods.dev/instructions/mouse-lin/finesse-skill/copilot-instructions"><img src="https://agentmods.dev/badge/instructions/mouse-lin/finesse-skill/copilot-instructions.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 | $0.01996 | $0.01996 |
| Opus 5 | $0.00998 | $0.00998 |
| Sonnet 5 | $0.00399 | $0.00399 |
| Haiku 4.5 | $0.00200 | $0.00200 |
Grade A, and why
finesse-skill copilot-instructions.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 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Copilot Instructions: finesse Design Standard
GitHub Copilot reads this file automatically. These rules replace generic AI design defaults with finesse's never-cheap, high-craft standards.
Core Directive
You are generating UI for a finesse project. finesse routes by register:
- brand (landing pages, hero sites, portfolios) — spectacle + soul + first impression
- product (dashboards, analytics, admin) — clarity + density + usability
- commerce (PDP, PLP, cart, checkout) — a hybrid; route to one of the two above by what the page does
- h5 (H5 / 移动端页面 / 活动页 / app 原型 / mobile PDP) — a page that only ever lives on a phone. A container register: fix the frame first, then wrap one of the three above. See rule 9.
All are never cheap. The premium substrate and anti-slop rules apply to all.
The 9 Non-Negotiables
- Read the brief first, and assert a direction the user can actually veto. Before any code, output
Design Read: {industry} · {soul} · register={brand|product|commerce|h5} · SPECTACLE={n}, then — on its own line —You'll see:a plain description of what renders (color, whether anything moves, type size, structure; no jargon, no library names), then two named likely objections (Not right? Most likely: ① … ② …). Stop and wait. The coordinate line is for you;You'll see:is the only line a non-designer can answer, and a gate only "go ahead" can pass is not a gate. Also name the lazy default aesthetic and how you are beating it — at two altitudes: the category default ("coffee → beige+brass") and the obvious anti-reference ("AI tool that's not SaaS-cream → editorial-typographic"). Avoiding the first reflex but landing on the second is still slop.- Before saying "done," check the built page against that
You'll see:line clause by clause. Anything that quietly became something else (a moving hero degraded to a static image, a dark theme drifting light) is a P0 — every other rule can pass while the page stops being what the user approved. - Name the image slots in that same block, and then wait. Add an
Images:line: how many pictures the page wants, what each depicts, and where they would come from in this session (check your actual tools first — image-gen available? network fetch only? neither?). Judge the slots off the skeleton, not the industry — a hero, a gallery, a PDP shot, an H5 cover or scene, even a dashboard's empty state or avatar row. Food/hotel/fashion briefs are the obvious cases, not the boundary; if the page truly wants none, writenoneand one clause of why. This is an offer the user answers — never a licence to act on it. Do not generate images and do not download or hotlink stock because you judged the brief implied them: both spend the user's budget or pull from the open network, and only an explicit yes to this question authorizes either. Silence, or approval of something else, is a no. Protocol:skills/finesse-ui/references/asset-sourcing.md. - Explain in observable terms, always. Gloss any term the user didn't use first — eyebrow (the small all-caps line above a headline), grain (a nearly invisible speckle so flat color doesn't read plastic). Never emit internal bookkeeping (five-axis coordinates,
differs on E + C + A, reference filenames).
- Before saying "done," check the built page against that
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 · 57 lines · 1,996 tokens per session scan A fbf9156441ba
finesse-skill copilot-instructions.md is an instructions file published in the GitHub repository mouse-lin/finesse-skill (537 stars, last pushed 1mo ago), licensed MIT. It adds 1,996 tokens to every session, about $0.0100 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 instructions, from other repositories
solidworks-api-skill CLAUDE.md
Instructions for guoleizhen717/solidworks-api-skill, covering solidworks api skill repository, repository purpose, skill scope and project conventions.
q3ds-solidworks-mcp AGENTS.md
AGENTS.md instructions for zarcherlot/q3ds-solidworks-mcp, covering codex repository guidance, architecture, drawing planning and verification.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.