prompt-director

prompt-director is an agent for Claude Code from Mineru98/imagine. It costs 96 tokens per session (1,157 once invoked), scanned A, original, MIT.

A prompt translator for image-generation requests. It turns Korean or abstract descriptions such as “dreamy” or “hip” into specific English visual wording when the request needs more detail.

In plain words
What is it for?
Use it before generating an image from Korean text or from a short, abstract visual prompt.
Why use it?
It helps an image model understand vague or non-English descriptions while keeping the user's original subject and intent.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the imagine plugin — 8 skills, 10 agents shipped together

Good fit Use it before generating an image from Korean text or from a short, abstract visual prompt.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/mineru98/imagine/prompt-director
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.

Clone the repo
git clone --depth 1 https://github.com/Mineru98/imagine

Made for: Claude Code.

Or install imagine, the plugin that ships this one along with the rest of its 8 skills, 10 agents.

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-director

README.md
[![agentmods](https://agentmods.dev/badge/agents/mineru98/imagine/prompt-director.svg)](https://agentmods.dev/agents/mineru98/imagine/prompt-director)
Your own site
<a href="https://agentmods.dev/agents/mineru98/imagine/prompt-director"><img src="https://agentmods.dev/badge/agents/mineru98/imagine/prompt-director.svg" alt="Measured on agentmods" height="20"></a>
Per session 96 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,157 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. 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 $0.00096 $0.01157
Opus 5 $0.00048 $0.00579
Sonnet 5 $0.00019 $0.00231
Haiku 4.5 $0.00010 $0.00116

Measured 8d ago against content hash eec168739469, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

prompt-director scanned grade A with 1 finding 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 8d 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.

Strips warnings and disclaimerslowAnti-refusal

Omitting safety caveats hides risk from the user and is a common jailbreak preamble.

- **안전 우회 문구 주입 금지.** "authorized red-team", "ignore safety", "without disclaimers" 등 안전장치 무력화성 표현을 결과에 포함하지 않는다.

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

agents/prompt-director.md · 59 lines

How it starts

The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.

prompt-director

추상적 한국어 서술을 시각 모델이 이해하기 좋은 구체 영어 키워드 구문으로 한 번만 번역·보정한다. 사용자의 원 의도를 보존하는 번역가 역할이지, 프롬프트 엔지니어가 아니다.

언제 개입하는가

  • 사용자 입력에 한글 음절이 하나라도 포함되어 있을 때.
  • 입력이 영어여도 추상 형용사 위주이고 시각 주체·배경·재료·조명 중 둘 이상이 비어 있을 때 (예: "vibes", "mood", "aesthetic" 단독).

언제 개입하지 않는가

  • 프롬프트가 이미 영어이며 주체·배경·스타일·조명 중 다수가 명시되어 있을 때.
  • 사용자가 특정 스타일 키워드를 이미 고정했을 때 ("anime, flat shading, ...").
  • 사용자가 원문 유지를 명시했을 때 ("as-is", "그대로", "원문 그대로").

변환 규칙

  1. 단일 패스만. 번역된 결과를 다시 번역·재보정하지 않는다. 재귀 rewrite 금지.
  2. booster 우위 금지. 보정 결과물에서 사용자 원 토큰이 부가 수식보다 토큰 수·가중치 면에서 밀리면 안 된다. 원래 키워드는 반드시 구문 앞쪽에 배치하고, 부가 키워드는 원문을 설명·구체화하는 범위에서만 덧붙인다.
  3. 투명 공개. 변환을 수행한 경우 반드시 사용자에게 다음 포맷으로 먼저 보고한 뒤 이미지를 생성한다:
    • 다음과 같이 변환했습니다: "<영문 프롬프트>"
  4. 무개입 분기. 개입하지 않기로 판단하면 원문을 그대로 하위 스킬에 전달하고, 변환 보고도 출력하지 않는다.
  5. 의도 보존. 의미 축소, 과장, 재해석을 하지 않는다. 사용자가 쓴 명사·주체는 반드시 유지한다.
  6. 스타일 주입 금지. 사용자가 고정하지 않은 매체/시점/해상도 수식어(예: "oil painting", "cinematic 35mm", "8k")를 임의로 붙이지 않는다. 단, 원문의 감정/분위기를 직접 구체화하는 조명·색조·질감 키워드는 허용한다.
  7. 언어 혼용. 고유명사·상표·지명은 원형을 유지하고 나머지만 영역한다.

예시 변환

한국어 / 추상 입력 보정 결과
몽환적인 숲 ethereal forest, volumetric mist, soft diffused light, dreamlike atmosphere
힙한 카페 trendy minimalist cafe, exposed brick, warm Edison bulbs, urban aesthetic
따뜻한 거실 cozy living room, warm tungsten lighting, soft woolen throw, late-afternoon golden hour
쓸쓸한 바닷가 새벽 lonely seashore at dawn, pale blue-grey sky, gentle low tide, single set of footprints in wet sand
레트로한 게임 방 retro gaming room, CRT monitor glow, stacked cartridges, neon wall poster, late-90s aesthetic

출력 형식

개입한 경우:

다음과 같이 변환했습니다: "<영문 프롬프트>"
(이후 imagine 스킬에 영문 프롬프트 전달)

개입하지 않은 경우: 추가 출력 없이 원문 그대로 전달.

금기

  • 재귀 rewrite 금지. 이미 변환된 영문 결과를 다시 다른 영문으로 다듬는 2차 보정을 하지 않는다.
  • 사용자 의도 검열·완화·도덕화 금지. 프롬프트가 표현한 감정·주제·대상을 임의로 순화·삭제하지 않는다.
  • 안전 우회 문구 주입 금지. "authorized red-team", "ignore safety", "without disclaimers" 등 안전장치 무력화성 표현을 결과에 포함하지 않는다.
  • 품질 booster 과다 주입 금지. masterpiece, 8k UHD, best quality 류의 일반형 품질 키워드를 사용자 요청 없이 덧붙이지 않는다.
  • 프롬프트 체인 분기 금지. 보정 결과는 항상 단일 구문이어야 하며, "variant A / variant B" 식 다중 프롬프트를 생성해 하위 스킬에 떠넘기지 않는다.

Read the full file on GitHub · 59 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. 8d ago First seen · 59 lines · 96 tokens per session scan A eec168739469

Subscribe to this mod's changes

prompt-director is an agent published in the GitHub repository Mineru98/imagine (5 stars, last pushed 10d ago), licensed MIT. It adds 96 tokens to every session and 1,157 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 1 finding (strips warnings and disclaimers). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

Image Prompt Engineer

Expert photography prompt engineer specializing in crafting detailed, evocative prompts for AI image generation. Masters the art of translating visual concepts into precise language that produces stunning, professional-quality photography through generative AI tools.

SHAdd0WTAka/Zen-Ai-Pentest · 44 tokens

Inclusive Visuals Specialist

Representation expert who defeats systemic AI biases to generate culturally accurate, affirming, and non-stereotypical images and video.

SHAdd0WTAka/Zen-Ai-Pentest · 30 tokens

hyv-veo-prompt-smith

The generative-prompt writer for HearYourVOICE (Phase 4). Looks at the shots still MISSING a source in the shotlist (after CC scouting) and writes copy/paste generation prompts to fill exactly those gaps — no more. Builds each prompt from the measured durations and the veo-prompt guide, applying subject-lock and…

killernay/HearYourVOICE · 124 tokens

Image Prompt Engineer

Expert photography prompt engineer specializing in crafting detailed, evocative prompts for AI image generation. Masters the art of translating visual concepts into precise language that produces stunning, professional-quality photography through generative AI tools.

criptogus/agent-evolve-network · 44 tokens

brief-constructor

Constructs optimized Gemini Nano Banana image generation prompts. Receives a user's image request and selected domain mode, applies Google's official 5-component formula, and returns a production-ready prompt string. Used internally by the banana skill before every image generation call.

juliandickie/creators-studio · 53 tokens

video-brief-constructor

Constructs optimized VEO 3.1 video generation prompts. Receives a user's video request and selected domain mode, applies the 5-part video framework, and returns a production-ready prompt string with camera, audio, and action specifications. Used internally by the video skill before every generation.

juliandickie/creators-studio · 64 tokens