contribute
Share improvements - issues, PRs, skills, templates, and videos you made
Install / Use
npx skills add digitalsamba/claude-code-video-toolkitInstalls into whichever agent you are using.
Claude Commands
Claude Code slash commands
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Our assessment of contribute
contribute scores 91/100 on our quality scale, 992nd of 4,702 Development & Engineering skills we index (top 22%).
Its Claude Commands is 14 KB long, well organised into 68 sections with 25 code examples: a thorough specification that gives an agent plenty to work with.
With 2,544 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated 5 days ago, so contribute is actively maintained.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit ā read the skill file before letting an agent act on it.
contribute compared with similar skills
All 4 of these similar skills score higher than contribute; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| contribute (this skill)by digitalsamba | 91 | 2.5k | 5d ago | Claude Commands |
| claude-memby thedotmack | 100 | 117.9k | today | CLAUDE.md |
| taste-skillby Leonxlnx | 100 | 117.6k | 9d ago | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 112.7k | 9d ago | CLAUDE.md |
| Understand-Anythingby Egonex-AI | 100 | 93.0k | 2d ago | CLAUDE.md |
Frequently asked questions
- How do I install contribute?
- Run
npx skills add digitalsamba/claude-code-video-toolkit. The install tabs above show the steps for each supported agent. - Which AI agents does contribute work with?
- It is written for Claude Code, as a Claude Commands file. Other agents that read the same format can often use it too.
- Is contribute safe to use?
- It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is contribute still maintained?
- The repository was last updated 5 days ago, so contribute is actively maintained.
Skill content
View source on GitHubdescription: Share improvements - issues, PRs, skills, templates, and videos you made
Contribute
Help users contribute improvements back to the toolkit without sharing their private project work.
Entry Point
When invoked, determine the contribution type:
How would you like to contribute?
1. **Report an issue** - Bug, missing feature, or question
2. **Submit an improvement** - PR for command, skill, template, or docs
3. **Share a new skill** - Package domain knowledge you've developed
4. **Share a template** - Generalize a project into a reusable template
5. **Share an example project** - Showcase a real project (configs + scripts, no large media)
6. **Share a video you made** - Post it in Show and tell (a link and how you made it, no project files)
Which would you like to do?
Option 1: Report an Issue
I'll help you create a GitHub issue.
What type of issue?
1. š Bug report
2. š” Feature request
3. š Documentation improvement
4. ā Question
Describe the issue:
Then generate issue with gh issue create:
gh issue create \
--repo digitalsamba/claude-code-video-toolkit \
--title "Bug: description" \
--body "## Description
[User's description]
## Steps to Reproduce
[If bug]
## Expected Behavior
[If bug]
---
Reported via /contribute command"
Option 2: Submit an Improvement
Step 1: Verify Safety
Let me check what would be shared...
ā
Safe to share:
.claude/commands/video.md (modified)
.claude/skills/ffmpeg/SKILL.md (modified)
ā Ignored (stays private):
projects/ (your video work)
assets/voices/ (your voice samples)
.env (secrets)
Does this look right?
Step 2: Create Branch & Commit
# Show current status
git status
# Create branch
git checkout -b improve/short-description
# Stage toolkit files only
git add .claude/ templates/ lib/ docs/ _internal/ tools/ brands/default/
# Commit with description
git commit -m "$(cat <<'EOF'
Improve: short description
Longer description of what changed and why.
š¤ Generated with Claude Code
EOF
)"
Step 3: Create PR
gh pr create \
--title "Improve: short description" \
--body "$(cat <<'EOF'
## Summary
What this PR improves.
## Changes
- Change 1
- Change 2
## Testing
How to test the improvement.
---
š¤ Generated with Claude Code
EOF
)"
Step 4: Confirmation
PR created! š
View at: [PR URL]
Your projects/ directory was NOT included.
Only toolkit improvements were shared.
After merge, you can update your local copy:
git checkout main
git pull
Option 3: Share a Skill
Step 1: Identify Skill
Which skill would you like to share?
Found in .claude/skills/:
1. ffmpeg (beta) - already in toolkit
2. my-new-skill (draft) - your addition
Or describe a new skill to create.
Step 2: Prepare Skill
Check skill is ready:
- [ ] SKILL.md has clear description
- [ ] Examples are tested and work
- [ ] No project-specific paths or secrets
- [ ] Status is at least "beta"
Step 3: Create PR
Same as Option 2, but focused on .claude/skills/{name}/
Option 4: Share a Template
Step 0: Is it actually a template? (ask this first)
Most "new template" ideas are really a brand profile. Ask before anything else, and take the answer seriously ā this is the single most common wrong turn in this command:
What makes your video different from the closest existing template?
a. Colors, fonts, logo, narrator voice
b. Scene pacing / timing values
c. Different scenes, different structure, different composition code
If (a) or (b) only -> you want a BRAND PROFILE, not a template.
If (a) or (b): stop here and run /brand instead. Templates read their
palette, fonts and voice from a brand (voice.brand in config.json points at
brands/<name>/voice.json), and timing values like lead/tail/xfade live
in your project's own config.json. Copying a template to change those produces
a near-identical fork that stops inheriting fixes to the original ā worse for
you than for the toolkit.
A brand profile does not need to be committed to this repo to work. Keep it local unless it is genuinely useful to other people.
Only continue below if (c) ā the composition code itself differs.
Step 1: Identify Template
Templates are generalized from projects.
Do you have a project you'd like to turn into a template?
Or an existing template in templates/ to share?
Step 2: Generalize (if from project)
To turn a project into a shareable template:
1. Copy to templates/:
cp -r projects/my-video templates/my-template
2. Remove project-specific content:
- Clear demo videos from public/demos/
- Reset config to example values
- Remove generated audio
- Update README with generic instructions
3. Verify it works:
cd templates/my-template
npm install
npm run studio
4. Check it is not a near-duplicate ā diff every file against the
template you started from:
for f in <new-template>/*; do
diff -q "<base-template>/$(basename "$f")" "$f" 2>/dev/null
done
Identical composition code (build.py, gen_captions.py, src/) means this
should be a brand profile plus project config, not a template. Go back
to Step 0.
Ready to proceed?
Step 3: Create PR
Same as Option 2, but focused on templates/{name}/
Option 5: Share an Example Project
Example projects showcase real-world usage. They include configs, scripts, and documentation but NOT large media files.
Step 1: List Available Projects
Found in projects/:
1. digital-samba-skill-demo
Marketing video for DS Claude Code skill
Template: product-demo style
2. sprint-review-cho-oyu
iOS sprint review for Digital Samba Mobile
Template: sprint-review
Which project would you like to share as an example?
Step 1b: Contributor Recognition
We'd like to credit you for sharing this example!
Your name or organization: ___
Website URL (optional, for backlink): ___
This will appear in:
- examples/README.md (Contributors table)
- The example's own README.md
- Project's CLAUDE.md
Leave blank to contribute anonymously.
Step 2: Prepare for Sharing
Preparing: digital-samba-skill-demo
I'll copy to examples/ and remove large media files.
The following will be INCLUDED:
ā
Configs and source code
- remotion/src/config/*.ts
- remotion/src/components/**
- remotion/package.json
ā
Documentation
- VIDEO-SPEC.md
- VOICEOVER-SCRIPT.md
- CLAUDE.md
- PROJECT-STATUS.md
ā
Small assets (logos, icons)
- remotion/public/images/*.png
- remotion/public/images/*.svg
The following will be EXCLUDED:
ā Large media (gitignored)
- remotion/public/demos/*.mp4
- remotion/public/audio/*.mp3
- node_modules/
- .venv/
- out/
ā Secrets
- .env
Proceed?
Step 3: Copy and Clean
# Copy project to examples
cp -r projects/{name} examples/{name}
# Remove large media (will be gitignored anyway, but clean up)
rm -rf examples/{name}/remotion/public/demos/*.mp4 2>/dev/null
rm -rf examples/{name}/remotion/public/audio/*.mp3 2>/dev/null
rm -rf examples/{name}/remotion/node_modules 2>/dev/null
rm -rf examples/{name}/remotion/.remotion 2>/dev/null
rm -rf examples/{name}/remotion/out 2>/dev/null
rm -rf examples/{name}/.venv 2>/dev/null
rm -f examples/{name}/.env 2>/dev/null
rm -f examples/{name}/.DS_Store 2>/dev/null
Step 4: Create ASSETS-NEEDED.md
Generate documentation of what media assets are needed:
# Assets Needed
To run this example, you'll need to create/provide these assets:
## Demo Videos
| File | Duration | Description |
|------|----------|-------------|
| `public/demos/skill-install.mp4` | ~15s | Screen recording of skill installation |
| `public/demos/app-demo.mp4` | ~30s | Browser recording of app flow |
## Audio
| File | Duration | Description |
|------|----------|-------------|
| `public/audio/voiceover.mp3` | ~2:30 | Generated from VOICEOVER-SCRIPT.md |
| `public/audio/background-music.mp3` | ~3:00 | Subtle tech ambient |
## How to Create
### Demo recordings
Use `/record-demo` command or Playwright scripts
### Voiceover
```bash
uv run tools/voiceover.py --script VOICEOVER-SCRIPT.md --output public/audio/voiceover.mp3
Background music
uv run tools/music.py --prompt "subtle tech ambient" --duration 180 --output public/audio/background-music.mp3
### Step 5: Update Example README
Ensure the example has a clear README.md with contributor credit:
```markdown
# Example: {name}
{Description from VIDEO-SPEC.md}
> Contributed by [{contributor_name}]({contributor_url})
## What This Demonstrates
- {Key feature 1}
- {Key feature 2}
- {Key feature 3}
## Quick Start
```bash
# Copy to your projects
cp -r examples/{name} projects/my-version
cd projects/my-version/remotion
# Install dependencies
npm install
# Create required assets (see ASSETS-NEEDED.md)
# ...
# Preview
npm run studio
# Render
npm run render
Structure
{Brief overview of project structure}
Assets
See ASSETS-NEEDED.md for required demo videos and audio files.
Contributed by {contributor_name}
Also update `examples/README.md` Contributors table with new entry.
### Step 6: Commit and PR
```bash
git add examples/{name}
git commit -m "Add example: {name}
{Description}
Note: Large media files not included. See ASSETS-NEEDED.md.
š¤ Generated with Claude Code"
gh pr create \
--title "Add example: {name}" \
--body "## Summary
Adds {name} as a showcase example.
## What it demonstrates
- {feature 1}
- {feature 2}
## Assets
Large media files are gitignored. Users create their own using:
- /record-demo for screen recordings
- /generate-voiceover for narration
- tools/music.py for background music
---
š¤ Generated with Claude Code"
Option 6: Share a Video You Made
For "look what I made". It posts in the repo's Show and tell Discussions category: a link and a few lines on how it was made. No project files leave the machine, and nothing is posted until the user has read the draft and said yes.
Step 1: Find the video and a link to it
Pick the project (projects/*/project.json), or ask for a video if it was made outside one. The post needs a link anyone can open:
project.jsonpublish.urlif/publishuploaded it to YouTube. Ifpublish.privacyStatusisprivate, nobody else can open it: ask them to make it unlisted or public in YouTube Studio first, or to use another link- otherwise ask for a link (YouTube, Vimeo, their own site)
- or they can drag the MP4 into the post on GitHub after it is created; leave a line saying where it goes
Never upload the video anywhere yourself.
Step 2: Gather how it was made (read only)
- From
project.json:title(orpublish.title),template,formats, thesessions[]summaries andrenderpaths. Not every template writes all of these; withoutrender, look for the newest MP4 in the project'sout/. - Length:
ffprobe -v error -show_entries format=duration -of csv=p=0 <render> - Tools used and cost: the lines of
logs/media-spend.log(JSON lines:tool,cost_usd,args,job_id) whoseargsmention the project's folder or whosejob_idis the project's job. Sumcost_usdand list the tools. If nothing matches, leave the cost out rather than guess. - What fought them: the sessions in
project.jsonoften say (a retake, a QC warning, a fix). Ask the user too: "Anything that broke, was slow or took a few tries?" This is the part the toolkit learns most from.
Step 3: Draft the post
**[<Title>](<link>)**, <length>. <One or two lines: what it is and who it's for.>
- How: <template, skills and t
Truncated for display ā read the full file on GitHub.
Related Skills
claude-mem
117.9kPersistent Context Across Sessions for Every Agent ā Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
taste-skill
117.6kAgent skills that give coding agents taste: art direction, reference images and a browser check instead of generic AI pages
Agent-Reach
112.7kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu ā one CLI, zero API fees.
Understand-Anything
93.0kGraphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit ā see the Safety scan above for what the skill file itself contains.
