Skip to content
Agentmart
Submit a tool

research-visuals

Research visual references and build annotated moodboards for visual creative work across media.

Skill by JoseArron

IndexedMIT★ 2
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/josearron-research-visuals

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Research visual references and build annotated moodboards for visual creative work across media. Use when choosing an aesthetic, comparing art territories, or investigating visual identity references.

Use cases: research

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/research-visuals

What's inside

Skill name
research-visuals
Description
Research visual references and build annotated moodboards for visual creative work across media. Use when choosing an aesthetic, comparing art territories, or investigating visual identity references.
License
MIT
Instructions
34 lines in SKILL.md · view on GitHub

Files (2)

  • SKILL.md
  • openai.yaml · agents/

SKILL.md (first lines)

# Research visuals

Find references that help the user make a visual decision. An annotated board explains what to borrow from each reference and what to leave behind.

## Start with the decision

Read the project's selected art contract, identity notes, owner feedback, and relevant experiments first. Separate selected direction, candidate direction, and rejected work. A convenient preview or recent file does not supersede an explicit decision.

Identify the visual decision: form, imagery, material, typography, composition, setting, or the whole identity. Use the audience, intended feeling, output medium, and production constraints already supplied. Ask only for a missing choice that changes the research.

## Collect usable evidence

- Choose reference territories from the brief's subject, audience, medium, and desired traits. Prefer creators' portfolios, official work or product pages, press kits, studio breakdowns, museum collections, and original publications. Trace reposted images to their creator when practical.
- Inspect the actual visual before describing it. If you can access only text, label that limit; do not invent a visual analysis.
- Record creator, work, source URL or local path, and the specific image, screen, or timestamp examined. For live references, verify the link.
- Describe visible traits: silhouette, proportion, contour, value grouping, palette relationships, material, lighting, type hierarchy, framing, and density. Separate observation from your proposed interpretation.
- Separate subject matter from visual treatment. Carry only the traits relevant to the request; a reference's setting, objects, or culture need not become the output's subject.
- For cultural references, identify the place, community, object, or practice represented. Verify language and heritage claims rather than treating a national label as a visual style.

## Build a useful comparison

Group references into distinct directions when the user is exploring. Change the underlying visual grammar, not just the palette. Each direction needs a rationale, useful traits, traits to exclude, and production implications.

When a board helps, use an editable HTML page or the requested format with labeled reference images and source links. Keep originals distinguishable from your generated studies. Use permitted embeds, supplied local images, or links; do not present a generated imitation as source evidence. If the request is research only, a cited comparison can be enough.

For each reference, answer: **What does this help us decide?** Omit attractive images with no role in that decision. Note when references contradict one another instead of implying that all their traits fit together.

## Deliver

Give a supported recommendation and annotated reference set. Mark unresolved choices as candidates. Save durable research in the project's established location, preferably `{agentRoot}/research/`, when useful. Follow the configured agent root; a one-off board needs no setup ceremony.

When the user selects a direction, use `define-art-direction` to turn it into production rules. Load its entrypoint through the harness's skill mechanism, or read `SKILL.md` directly when no mechanism is available. Research does not itself approve a direction.

From JoseArron/skills, MIT.

More from JoseArron/skills

Signals

Stars
2
Forks
0
Open issues
0
Last commit (this folder)
Oct 11, 2026 (18 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from JoseArron

Cite this page

research-visuals skill. Agent Mart. https://agentmart.dev/skills/josearron-research-visuals (last updated 2026-10-24)