ralphy: Skill for Claude Code

.agents/skills/art-director/SKILL.md

art-director is a skill for Claude Code, Codex from alecs5am/ralphy. It costs 154 tokens per session (5,845 once invoked), scanned A, original, Apache-2.0.

A production playbook for turning an approved creative scenario into prompts and generated images, video, voice-over, or music. It also sets rules for keeping locations and characters visually consistent across scenes.

In plain words
What is it for?
Use it to create prompts, generate or regenerate media, try model variants, and manage reference images for multi-scene projects.
Why use it?
It reduces mismatched locations, characters, and visual styles between generated scenes. It also helps choose models, estimate costs, and decide when to regenerate an asset.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: positional $N argument; installed under .agents/ (shared by several agents); mentions AGENTS.md.

This is alecs5am/ralphy's own configuration. It tells Claude Code and Codex how to work on ralphy 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 ralphy configures →

Reuse

Borrowing it

Nothing to install: this file belongs to alecs5am/ralphy. 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/alecs5am/ralphy/main/.agents/skills/art-director/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/alecs5am/ralphy

Made for: Claude Code, Codex.

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 art-director

README.md
[![agentmods](https://agentmods.dev/badge/skills/alecs5am/ralphy/art-director/github.svg)](https://agentmods.dev/skills/alecs5am/ralphy/art-director)
Your own site
<a href="https://agentmods.dev/skills/alecs5am/ralphy/art-director"><img src="https://agentmods.dev/badge/skills/alecs5am/ralphy/art-director/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 art-director

Your own site · 80×15
<a href="https://agentmods.dev/skills/alecs5am/ralphy/art-director"><img src="https://agentmods.dev/badge/skills/alecs5am/ralphy/art-director.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 154 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,845 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. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 3 findings, up to high

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • high Anti-Refusal · line 116
    Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.
    Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
  • high Tool Misuse · line 167
    Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
    Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
  • medium Agent Snooping · line 117
    Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
    Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
How audits are shown
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 $0.00154 $0.05845
Opus 5 $0.00077 $0.02923
Sonnet 5 $0.00031 $0.01169
Haiku 4.5 $0.00015 $0.00585

Measured 9d ago against content hash b1d3162ccaac, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

art-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 9d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

> **STOP rule.** Every model call goes through `ralphy generate`. No raw `fetch` / `curl` / `bunx tsx` against a media API — gen-log + asset-manifest + cost rollup all depend on the CLI. AGENTS invariant #2.
.agents/skills/art-director/SKILL.md · 209 lines

How it starts

The opening of the file, as written. The whole thing — 209 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Art director playbook

Read this when: "generate prompts", "generate assets", "make images / video / VO / music", "regenerate scene-XX ", "try a different model", "A/B variant", "how much will it cost".

Anchor order discipline (every multi-scene project):

  1. Location-master-plate first — for any project where ≥2 scenes share a setting, generate the room / location plate as anchor #1, BEFORE any character or scene anchor. Pass the plate as --ref alongside character masters on every subsequent scene gen. Skipping this cost noski-people-001 $0.45 image-regen + 45 min user-feedback loop ("in every shot they're sitting on a different couch and I asked for the same one" — three different couches across three anchors). For ≥25-scene projects, the plate alone isn't enough — generate ≥3 unique anchor angles per recurring subject (location, hero character, hero product). Full discipline + CLI shape + worked example: art-director/location-plate.md. This is the single highest-leverage rule in this playbook.
  2. Character / persona masters second — one per cast member, each generated with the location plate as --ref. Pass both (location + character) on every downstream scene gen to lock identity + setting.
  3. Scene anchors third — scene-01 first, surfaced to user → wait → scene-02 → wait → … only batch 4-6 anchors at a time AFTER two solo gens land with user approval.
  4. i2v / video generation last — never i2v an unapproved scene anchor.

Photoreal-human projects: read art-director/photoreal-humans.md before drafting prompts — TV-commercial register (Tom-Ford / chiaroscuro / marble) is the wrong default for natural-feeling UGC; use Sony A7 IV + Sigma 35/85mm + Kodak Portra 400 still-photo register instead. Venom-bodywash-001 burned ~$3 on this miscalibration.

Aesthetic-lock checkpoint (HARD gate, applies before anchor fan-out): after the location-master-plate (anchor #1) and the character master(s) (anchor #2) are approved by the user, the agent MUST (a) name the register in one phrase ("still-photo candid documentary" / "Old-Spice high-key commercial" / "liminal-spaces analog-horror" / "CGI translucent specimen" / etc. — taxonomy at the top of photoreal-humans.md), (b) run the character-fit.md check (clean mascot + gritty register? pick reinterpret / distressed-variant / shift-register BEFORE fan-out), (c) write the one-line aesthetic-lock string into STORYBOARD.md, and (d) generate one cover per register and surface it to the user. Refuse to fan out scene anchors until the cover is approved. Skipping this is the documented cause of ralphy-vs-higgsfield-001's three-register shipping defect (Lesson #7) and ralphy-carousel-001's clean-ghost-in-punk drift (postmortem #2). When in doubt about which register applies, run ralphy ref pull <one-canonical-url> --frames and READ the frames — do NOT scrape-summarize (biofix-hypnic-en-001 defect class).

Model drift handling: read art-director/regeneration.mdone retry max on a kling/seedance prompt that misses; then redesign the scene, don't fight model basins. Glitter-cream-001 lost 2× $0.42 fighting "jar near cheek → powder compact" drift across 3 retries.

Between "scenario approved" and "assets on disk for the editor" — that's my zone. Prompt engineering, API orchestration, single-slot regeneration, A/B variants, cost discipline. Never invent model-id from memory — always cross-check MODELS.md.

STOP rule. Every model call goes through ralphy generate. No raw fetch / curl / bunx tsx against a media API — gen-log + asset-manifest + cost rollup all depend on the CLI. AGENTS invariant #2.

Read the full file on GitHub · 209 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. 9d ago First seen · 209 lines · 154 tokens per session scan A b1d3162ccaac

Subscribe to this mod's changes

art-director is a skill published in the GitHub repository alecs5am/ralphy (132 stars, last pushed 14d ago), licensed Apache-2.0. It adds 154 tokens to every session and 5,845 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

create-art-prompt

Fabric pattern: createartprompt.

bdmorin/the-no-shop · 11 tokens

segment-anything-model

Foundation model for image segmentation with zero-shot transfer. Use when you need to segment any object in images using points, boxes, or masks as prompts, or automatically generate all object masks in an image.

synthetic-sciences/openscience · 45 tokens

blip-2-vision-language

Vision-language pre-training framework bridging frozen image encoders and LLMs. Use when you need image captioning, visual question answering, image-text retrieval, or multimodal chat with state-of-the-art zero-shot performance.

synthetic-sciences/openscience · 52 tokens

sc:image-gen-prompting

Craft effective text-to-image prompts for AI generators (Higgsfield, Midjourney, DALL·E/GPT-Image, Stable Diffusion, Flux). Subject + style + composition + lighting structure, negative prompts, aspect ratios, model-specific tips, and Hebrew→English prompt translation. Activate for: generate image, AI art, product…

squadcodercom/squadcoder · 95 tokens

histolab

Lightweight WSI tile extraction and preprocessing. Use for basic slide processing tissue detection, tile extraction, stain normalization for H&E images. Best for simple pipelines, dataset preparation, quick tile-based analysis. For advanced spatial proteomics, multiplexed imaging, or deep learning pipelines use pathml.

synthetic-sciences/openscience · 62 tokens

shap

Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Use this skill when explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing…

synthetic-sciences/openscience · 109 tokens