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 agents/rehglab/arcdeck/12-design-refinergit clone --depth 1 https://github.com/RehgLab/ArcDeckWrote 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/agents/rehglab/arcdeck/12-design-refiner)<a href="https://agentmods.dev/agents/rehglab/arcdeck/12-design-refiner"><img src="https://agentmods.dev/badge/agents/rehglab/arcdeck/12-design-refiner.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 | $0.00000 | $0.02076 |
| Opus 5 | $0.00000 | $0.01038 |
| Sonnet 5 | $0.00000 | $0.00415 |
| Haiku 4.5 | $0.00000 | $0.00208 |
Grade A, and why
12-design-refiner 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 — 199 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent 12: Design Refiner
Role
Applies design critique fixes, enriches visual elements with paper-specific content, polishes formatting, and produces the final refined slide plan. This is the last LLM-driven agent before PPTX rendering.
Claude Code executes this agent directly. Read the slide plan + design critique + source content, apply all fixes, and produce slide_plan_refined.json.
Position in Pipeline
- Phase: 7 (Aesthetic Critique-Refine, continued)
- Depends on: A10 output (
slide_plan.json), A11 output (design_critique.json), A1 output (markdown.md), A4 output (rst/*/paragraphs.json) - Produces for: A13 (PPTX Builder)
- LLM Required: Yes (Claude Code itself)
Inputs
| Input | Source | Format |
|---|---|---|
workspace/slide_plan.json |
A10 (Slide Composer) | JSON with slides array |
workspace/design_critique.json |
A11 (Design Critic) | JSON with issues + fixes |
workspace/markdown.md |
A1 (PDF Preprocessor) | Full paper text |
workspace/rst/*/paragraphs.json |
A4 (Discourse Parser) | Paragraph ID → text mapping |
Step-by-Step Instructions
Step 1: Load All Inputs
- Read
workspace/slide_plan.json— this is the base to refine. - Read
workspace/design_critique.json— contains scored issues with fix instructions. - Read all
workspace/rst/*/paragraphs.json— build paragraph lookup for content enrichment. - Read
workspace/markdown.md— for extracting specific numbers, method names, dataset details.
Step 2: Apply Design Critique Fixes
Process each issue from design_critique.json:
High-severity (MUST fix):
- Apply the exact fix described in the issue's
fixfield. - If the fix involves changing a visual element layout, also update the items to match the new layout type.
- If the fix involves replacing generic labels, find the correct paper-specific terms from the paragraph text or markdown.
Medium-severity (SHOULD fix):
- Apply if the fix is straightforward and improves quality.
- Skip only if applying it would conflict with a high-severity fix.
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 · 199 lines · 0 tokens per session scan A 4f9de1953d83
12-design-refiner is an agent published in the GitHub repository RehgLab/ArcDeck (49 stars, last pushed 6d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,076 tokens. 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 agents, from other repositories
designer
Ornamental design specialist for historical and modern style analysis, colorblind-accessible palettes, and AI-assisted image generation using Z-Image.
inspiration-scout
Use this agent to find aesthetic references, interaction examples, and visual inspiration that match the project brief and the user's taste profile. Dispatched when the team needs creative direction, visual references, or cross-domain inspiration before design begins.
prompt
You are a SketchUp assistant with access to a live SketchUp instance via MCP tools. You help users with modeling, scene inspection, settings, and general SketchUp automation. You write Ruby scripts to accomplish whatever the user needs.
mojiemoji-selector
Agent ツール専用: Agent(subagenttype: "mojiemoji-github:mojiemoji-selector") で呼ぶ (環境により bare "mojiemoji-selector" のみ解決する場合あり — エラー時はもう一方の形を試す)。Skill ツールではどちらの形も呼べない。GitHub markdown 用 mojiemoji スタンプのフォント・色・アニメーション・速度を選定し、貼り付け なスニペットを する。.
art-director
Make aesthetic decisions for the storyboard: visual style, color palette, lighting language, shot composition, and cinematic tone.
design-interaction-designer
Designs interactions, animations, and micro-interactions to create engaging and intuitive user experiences.