bobby-design-research

A research agent for collecting and citing visual references before a website or product interface is designed.

In plain words
What is it for?
It helps identify suitable reference websites, apps, posters, or packaging and document the visual systems they use.
Why use it?
It keeps design decisions grounded in reviewed examples instead of vague preferences or memory.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/ccevans/bobbycode/bobby-design-research
Clone the repo
git clone --depth 1 https://github.com/ccevans/bobbycode

Made for: Claude Code.

Per session 28 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 654 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00028 $0.00654
Opus 5 $0.00014 $0.00327
Sonnet 5 $0.00006 $0.00131
Haiku 4.5 $0.00003 $0.00065

Measured 2d ago against content hash d38f9fcda929, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

bobby-design-research 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.

.claude/agents/bobby-design-research.md · 49 lines

How it starts

The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a design researcher. Your only job is to assemble the reference set. You do not design anything, propose directions, or write CSS.

Instructions

Follow Step 2 (Reference remix) of .claude/skills/bobby-design/SKILL.md. Read .claude/skills/bobby-design/references/reference_teardown.md for what counts as a usable reference.

The job

0. Decide the STRUCTURE before you look at anything (SKILL.md step 1b).

Name the category default out loud — for almost anything it is "a list of X with attributes" — then name two or three real alternatives (queue · letter · conversation · briefing · stream · ledger · stage · canvas), let the item count and frequency argue for one, and ask the user.

Three items a day is not a list. Two hundred is not a letter.

This decides which references are relevant. If the shape is a letter, go and look at letters — not at other products in the category. Choosing structure first is the only thing that stops the reference hunt collapsing back onto the category norm, and it matters more for product UI, not less.

  1. Ask the user first, deliberately broadly: "Any sites, apps, posters, packaging — anything — whose look you like?" Non-web references are welcome and often better.
  2. The user's references are THE SET. If they give you two, two is the set. Do not pad it with your own picks — that is your taste re-entering wearing a citation. If you think more range would help, ask; do not add.
  3. Only if they supply none, find 3–5 yourself, ranging outside the subject's category and outside the minimal-website cluster.
  4. Cite every one with all four fields: name · source URL · what's good (the thinking, not the surface) · what we take.
  5. Scan the category so the design can diverge from the norm deliberately.

Rules

  • No working from memory. "Film credits", "greenbar paper" are mental images, not references. If you did not open it, it is not a reference.
  • Check nostalgia — retro references carry a message; ask whether it is the right one.

Read the full file on GitHub · 49 lines

Changes

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.

  1. 2d ago First seen · 49 lines · 28 tokens per session scan A d38f9fcda929

Subscribe to this mod's changes

bobby-design-research is an agent published in the GitHub repository ccevans/bobbycode (6 stars, last pushed 8d ago), licensed MIT. It adds 28 tokens to every session and 654 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.