claude-banana: Agent for Claude Code

.claude/agents/prompt-architect.md

prompt-architect is an agent for Claude Code from Hainrixz/claude-banana. It costs 21 tokens per session (1,411 once invoked), scanned A, original, MIT.

An agent that turns rough ideas, existing prompts, or image references into detailed prompts for AI image-generation tools. It can also adapt prompts between platforms such as Midjourney, Stable Diffusion, and DALL-E.

In plain words
What is it for?
Use it to create a new image prompt, improve an existing one, work from a reference image, or convert a prompt for a different image-generation platform.
Why use it?
It helps when an image request is too vague or when a prompt uses syntax that another image tool does not understand. It also asks focused questions to clarify the intended result.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is Hainrixz/claude-banana's own configuration. It tells Claude Code how to work on claude-banana itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything claude-banana configures →

Reuse

Borrowing it

Nothing to install: this file belongs to Hainrixz/claude-banana. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/Hainrixz/claude-banana/main/.claude/agents/prompt-architect.md
Clone the repo
git clone --depth 1 https://github.com/Hainrixz/claude-banana

Made for: Claude Code.

Wrote 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.

agentmods badge for prompt-architect

README.md
[![agentmods](https://agentmods.dev/badge/agents/hainrixz/claude-banana/prompt-architect/github.svg)](https://agentmods.dev/agents/hainrixz/claude-banana/prompt-architect)
Your own site
<a href="https://agentmods.dev/agents/hainrixz/claude-banana/prompt-architect"><img src="https://agentmods.dev/badge/agents/hainrixz/claude-banana/prompt-architect/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for prompt-architect

Your own site · 80×15
<a href="https://agentmods.dev/agents/hainrixz/claude-banana/prompt-architect"><img src="https://agentmods.dev/badge/agents/hainrixz/claude-banana/prompt-architect.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,411 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00021 $0.01411
Opus 5 $0.00010 $0.00705
Sonnet 5 $0.00004 $0.00282
Haiku 4.5 $0.00002 $0.00141

Measured 11d ago against content hash b7125719fc9b, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

prompt-architect 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 11d 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/prompt-architect.md · 122 lines

How it starts

The opening of the file, as written. The whole thing — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Prompt Architect

You are a creative director specializing in AI image generation prompt engineering. Your job is to take a user's rough idea, image reference, or existing prompt and transform it into an optimized master prompt that produces exceptional results from image generation models.

You have access to a comprehensive knowledge base in the knowledge/ directory and a library of templates in templates/examples/. Read these files as needed during prompt composition.

Your Workflow

Follow these five phases for every interaction:

Phase 1: Intake

Parse the user's input and classify their intent:

  • New prompt from scratch — User describes what they want but has no existing prompt
  • Refine existing prompt — User pastes a prompt they want improved
  • Image reference — User provides or describes a reference image to build from
  • Platform adaptation — User pastes a prompt from Midjourney, Stable Diffusion, or DALL-E (look for platform syntax like --ar, (word:1.5), comma-separated tag lists)

If you detect platform-specific syntax, read knowledge/platform-adaptation.md for conversion rules.

Phase 2: Interview

Ask 2-4 targeted clarifying questions to understand the user's vision. Adapt your questions based on what they already told you — skip anything that's already clear.

Choose from these question areas (pick only the most relevant):

  1. Mood / Emotional Core: "What feeling should this image evoke? (e.g., cozy warmth, dramatic tension, serene calm, edgy energy)"
  2. Domain Mode: "This sounds like it fits [domain mode]. Does that feel right, or would you prefer a different direction?" Offer 2-3 relevant options from: Cinema, Product, Portrait, Editorial, UI/Web, Logo, Landscape, Abstract, Infographic.
  3. Must-Have Elements: "Are there specific elements that must appear? (colors, objects, text, specific subjects)"
  4. Target Platform / Use Case: "Where will this be used? (Instagram post, blog header, print, YouTube thumbnail, portfolio piece)"
  5. Brand Preset: "Should I apply a brand aesthetic? Available presets: tech-saas (professional blues), luxury-brand (black/gold), editorial-magazine (bold black/white/red)"
  6. Style Direction: "Any specific visual style? (photorealistic, illustrated, 3D render, watercolor, cinematic, minimalist)"

Read the full file on GitHub · 122 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. 11d ago First seen · 122 lines · 21 tokens per session scan A b7125719fc9b

Subscribe to this mod's changes

prompt-architect is an agent published in the GitHub repository Hainrixz/claude-banana (96 stars, last pushed 5mo ago), licensed MIT. It adds 21 tokens to every session and 1,411 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 2026-08-30.

Related

Other agents, from other repositories