Handwritten Logo Reveal Codex Skill

Do not rely on it for

Narrated, multi-scene, or live-action video

Recovering authentic pen order from filled SVG outlines or raster pixels

Automatic rights clearance, identity verification, creative approval, publication, or deployment

The ZIP does not bundle HyperFrames, a browser, FFmpeg, third-party fonts, sound effects, or sample video

Review the package before enabling it.

Download the v1.0 ZIP and inspect its text files. For one project, create .agents/skills and extract the ZIP there. For user scope, extract it into $HOME/.agents/skills. Confirm codex-handwritten-logo-reveal/SKILL.md exists, restart Codex if needed, then use /skills to verify discovery.

Trigger

Use when the user asks Codex to create or revise a short, unnarrated handwritten logo, signature-style wordmark, luminous draw-on, or writing-head reveal from an SVG, raster mark, or exact text.

Do not trigger

Do not use for narrated explainers, multi-scene edits, live-action footage, prompt-only ideation, or an exact real signature that cannot be preserved from the supplied source.

Permissions

Read the supplied mark, reference media, and user-scoped project files; Create or edit files only inside the user-scoped video project; Run local profiling, media inspection, HyperFrames checks, snapshots, and Studio preview; Do not copy reference assets, publish, upload, or render before the stated approval boundary

Dependencies

A current HyperFrames CLI installation and its relevant agent skills; Node.js 18 or newer for scripts/profile-logo-input.mjs; FFmpeg and ffprobe for media and encoded-file inspection; A local browser supported by the installed HyperFrames version

This downloadable Codex skill turns the handwritten-logo method into a narrow reusable workflow. It activates for short, unnarrated signature-style logo or wordmark reveals, profiles the supplied source, chooses an honest reveal strategy, builds a deterministic HyperFrames project, and pauses at a checked Studio preview before rendering.

What the ZIP contains

README.md explains project-scoped and user-scoped installation.
SKILL.md defines the trigger, intake, production sequence, hard boundaries, and reporting contract.
scripts/profile-logo-input.mjs profiles exact text, SVG, or raster input and reports the recommended branch.
references/input-strategies.md distinguishes stroke drawing, filled-SVG masks, raster masks, and typeset masks.
references/quality-gates.md separates composition checks, Studio approval, rendering, and encoded-file QA.
assets/grow-ai-skills-example-brief.md supplies the exact 5.8-second example specification.
assets/behavior-test-record.md records the retained composition checks, approved render, and encoded-file readback.
agents/openai.yaml provides optional Codex interface metadata and a starter request.

Install for one project

Run from the target project root

mkdir -p .agents/skills
unzip codex-handwritten-logo-reveal-skill-v1-0.zip -d .agents/skills

Confirm that .agents/skills/codex-handwritten-logo-reveal/SKILL.md exists. Restart Codex if necessary and use /skills to confirm discovery.

Watch or download the retained Grow AI Skills behavior-test video. The MP4 is evidence for this specific run; it is not bundled inside the lightweight Skill ZIP.

Install for all projects

macOS or Linux

mkdir -p "$HOME/.agents/skills"
unzip codex-handwritten-logo-reveal-skill-v1-0.zip -d "$HOME/.agents/skills"

Windows PowerShell

New-Item -ItemType Directory -Force "$HOME\.agents\skills" | Out-Null
Expand-Archive .\codex-handwritten-logo-reveal-skill-v1-0.zip -DestinationPath "$HOME\.agents\skills" -Force

Start it explicitly

Grow AI Skills example

Use $codex-handwritten-logo-reveal to create a 5.8-second 1920x1080 reveal of the exact text “Grow AI Skills”. Use a black background, a white finished wordmark, and a magenta-to-cyan writing head. Preview first; do not render until I approve.

What it will not claim

The skill does not treat a filled SVG's outline order as real pen order, represent typeset text as a person's authentic signature, copy a visual reference's mark or audio, use remote runtime dependencies, or call a passing browser check a finished MP4. It also does not publish or deploy the result.

Verification boundary

The source package is structurally validated and its profiling script is tested against exact text, a local SVG, and an empty-input failure. The retained Grow AI Skills example supplies the behavior test for composition assembly, preview, approved rendering, and encoded-file inspection. Production deployment and destination publication remain separate states.