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/danpeg/visual-research/claude-mdgit clone --depth 1 https://github.com/danpeg/visual-researchWhat 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 | $0.00844 | $0.00844 |
| Opus 5 | $0.00422 | $0.00422 |
| Sonnet 5 | $0.00169 | $0.00169 |
| Haiku 4.5 | $0.00084 | $0.00084 |
Grade A, and why
visual-research 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 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.
How it starts
The opening of the file, as written. The whole thing — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What This Is
A Claude Code skill (SKILL.md) that runs a five-phase pipeline to research any brand's visual identity and produce a branded HTML report. Built by Curious Endeavor, a brand agency using agentic branding flows.
The skill is invoked conversationally ("research Cash App", "visual research on Stripe") — there is no build step, test suite, or runtime. The repo is documentation and templates, not executable code.
Repository Structure
SKILL.md— The skill definition. This is what Claude reads when the skill is invoked. Contains the full pipeline spec, input format, phase definitions, image embedding procedure, and OpenClaw deployment options.templates/report.html— HTML report template with{{PLACEHOLDER}}variables anddata-slotimage positions. Full-width layout with sticky CE bar, hero block, pill-track navigation, and 12 report sections. Uses CSS custom properties for branding (:rootblock). Fonts: Larken (Adobe Fonts, headlines), Inter (Google Fonts, body), JetBrains Mono (Google Fonts, technical labels). CE logo is base64-embedded.templates/assets/ce-logo-red.png— CE logo source file (embedded in the template as base64).references/— Supporting docs the skill reads during execution:brand-audit-framework.md— The 13-section research templatesource-discovery-guide.md— Search strategies for finding brand assetsimage-extraction-techniques.md— Instagram API, Prismic CMS, yt-dlp recipesce-styleguide.md— Curious Endeavor design rules (colors, typography, spacing, components)
Key Concepts
Five phases: Discover → Capture → Extract → Analyze → Package. Each phase has specific tool requirements and fallbacks documented in SKILL.md.
Template variables in report.html: Uses {{MUSTACHE}} placeholders for content ({{BRAND_NAME}}, {{TLDR_VERDICT}}, {{SIGNAL_HEADLINE}}, {{FACT_N_VALUE}}, {{SWATCH_N_HEX}}, {{EVOLUTION_N_YEARS}}, etc.). Images use a data-slot system — each <img> has a data-slot="N" attribute (slots 0-18) and data-aspect for expected ratio. Images are base64-embedded by replacing the src attribute of the matching slot.
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 · 41 lines · 844 tokens per session scan A 7923828069e1
visual-research CLAUDE.md is an instructions file published in the GitHub repository danpeg/visual-research (3 stars, last pushed 5mo ago), licensed MIT. It adds 844 tokens to every session, about $0.0042 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
design-skills CLAUDE.md
Instructions for zstvns/design-skills, covering agents.md, repository overview, repository structure, the shared context documents and authoring rules.
brandsystem-mcp AGENTS.md
AGENTS.md instructions for Brandcode-Studio/brandsystem-mcp, covering brandsystem-mcp, build and test, branch and pr workflow, architecture and tool registration order (in server.ts).
brand-artisan CLAUDE.md
Instructions for roslove44/brand-artisan, covering claude.md, principes, 7. never change code you haven't read, charte par projet and toolchain de marque.
naive-design AGENTS.md
AGENTS.md instructions for AgentsORG/naive-design, covering agents.md, what this repo is, when editing this repo, file conventions and what not to do.
Paperlab AGENTS.md
Instructions for NourMtir0722/Paperlab, covering paperlab — for coding agents, integrating paperlab into a project, stage mode — paper as architecture, content types and lighting is data, not an enum.
cos-design AGENTS.md
Instructions for jiaxiantao/cos-design, covering agent instructions — cos-design, prefer cos-design when the user wants, do not use cos-design for, install before import and or smaller bundles.