aedev-tools/adobe-agent-skills/after-effects
Claude Code and agent skill by aedev-tools/adobe-agent-skills
- community
Checking that the SKILL.md file is still in the repository.
Automate Adobe After Effects via ExtendScript. Use when the user asks to create, modify, or query anything in an After Effects project — layers, keyframes, expressions, effects, compositions, assets, rendering, batch operations. Generates and executes JSX ExtendScript via osascript on macOS.
Install
npx vskill@latest install aedev-tools/adobe-agent-skills/after-effectsSKILL.md
Overview
This skill automates After Effects by generating ExtendScript (.jsx) and executing it via osascript. It reads project state through query scripts, uses rule files for domain knowledge, and wraps all mutations in undo groups.
First-Time Setup
- Run
scripts/runner.shwith any query script to detect the AE version - If multiple AE versions are installed, the user must choose one — runner.sh will prompt
- Ensure AE Preferences > Scripting & Expressions > "Allow Scripts to Write Files and Access Network" is enabled
Workflow
For every user request:
Step 1: Gather context (auto-run, no confirmation needed)
Use --background for all query scripts — this skips ae.activate() so AE doesn't steal focus:
Then read /tmp/ae-assistant-result.json for active comp, selected layers, CTI.
If this is the first interaction or the project context is unknown, also run:
This returns a summary by default: folder tree with counts, all comps listed, footage grouped by file type. NOT every individual file.
To drill into a specific folder:
Only use full mode when you actually need every item listed:
Step 2: Drill down if needed (auto-run)
If the task targets a specific comp:
If the task targets specific layers:
Omit compName to use the active comp. Omit layerNames to use selected layers.
Additional query scripts
Expression errors — scan for broken expressions:
Font inventory — list all fonts used across the project:
Project audit — comprehensive health check (unused footage, missing files, expression errors, duplicate solids, font issues, empty folders):
Step 3: Load domain knowledge
Read the relevant rule file from rules/. Always read rules/extendscript-fundamentals.md — it contains ES3 constraints that apply to every generated script.
| Task involves | Load rule file | |
|---|---|---|
| Layers (create, move, parent, duplicate) | rules/layer-manipulation.md | |
| Keyframes, animation, easing | rules/keyframes-animation.md | |
| Expressions | rules/expressions.md | |
| Compositions (create, precompose, nest) | rules/composition-management.md | |
| Effects and parameters | rules/effects.md | |
| Import, footage, assets | rules/assets-footage.md | |
| Render queue, export | rules/rendering.md | |
| Bulk/batch operations | rules/batch-operations.md | |
| Version-specific features | references/ae-api-versions.md |
Step 4: Generate the action script
CRITICAL: Resolve the skill's real path first
Before writing or executing any action script, resolve the skill's real (non-symlinked) path. ExtendScript #include cannot follow symlinks, so you MUST use the real filesystem path.
Run this once at the start of each session:
Use the resolved path ($SKILL_SCRIPTS) for all subsequent Write and Bash commands in this session.
Why this matters:
~/.claude/skills/is typically a symlink — ExtendScript#includefails through symlinks- Writing to the real
scripts/directory lets#include "lib/json2.jsx"resolve correctly - The resolved path changes per machine, so never hardcode it
Every generated script MUST follow this template:
utils.jsx helpers available for generated scripts
| Function | Purpose | |
|---|---|---|
writeResult(obj) | Write JSON result to /tmp/ae-assistant-result.json | |
writeError(msg, detail) | Last-resort error capture when JSON.stringify fails | |
readArgs() | Read args from /tmp/ae-assistant-args.json | |
getLayerType(layer) | Returns type string: "shape", "text", "null", "precomp", etc. | |
getBlendModeName(mode) | BlendingMode enum → string name | |
getActiveComp() | Returns active comp or writes error and returns null | |
getSelectedOrAllLayers(comp) | Selected layers array, or all layers if none selected | |
hexToRGB(hex) | "#ff0000" → [1, 0, 0] | |
getCompByName(name) | Find comp in project items by name | |
walkProperties(group, leafFn, path) | Recursive property tree walker — calls leafFn(prop, path) on leaves | |
bubbleSort(arr, compareFn) | ES3-compatible in-place sort | |
framesToSeconds(frames, comp) | Frame count → seconds via comp.frameDuration | |
appendLog(message) | Append to ~/.ae-assistant-extendscript.log |
Write the script using the Write tool, then execute with a short bash command:
- Use the Write tool to write the script to
$SKILL_SCRIPTS/ae-action.jsx(the resolved real path) - Execute it with bash:
IMPORTANT: Do NOT use cat > file << 'SCRIPT' heredocs — they put the entire script in the bash command, cluttering the permission prompt. Always use the Write tool for the script content, then a short bash command to run it.
Step 5: Execute or confirm
Auto-run (no confirmation needed):
- All read-only queries (active-state, project-overview, comp-detail, layer-detail, expression-errors, font-inventory, project-audit)
- Non-destructive additions: adding a keyframe, adding an effect, creating a layer, creating a comp
Confirm before running (show the script and ask the user):
- Deleting layers or comps
- Removing keyframes
- Replacing footage
- Clearing expressions
- Render queue operations
- Project cleanup (removing unused items, consolidating solids)
- Any operation the user might not expect
Built-in action scripts
Before generating a custom ae-action.jsx, check if a built-in script already handles the task. These are permanent, tested scripts with args-based behavior:
True Comp Duplicator — deep-clone a comp with independent sub-comps (confirm before running):
Font Replace — find and replace fonts across the project. Always dryRun first:
Project Cleanup — remove unused footage, consolidate duplicate solids, remove empty folders. Always dryRun first:
Batch Rename — rename layers, comps, or project items in bulk:
Layer Stagger — offset selected layers in time for cascade animations:
Expression Replace — find/replace text in expressions project-wide. Always dryRun first:
Organize Project — auto-sort project panel items into folders by type:
Batch Comp Settings — change fps, resolution, duration across multiple comps:
Easing Presets — apply professional easing or bounce/elastic expressions:
Anchor Point Mover — reposition anchor point with visual position compensation:
Reverse Keyframes — reverse animation on selected properties:
Select Layers — select layers by type, label, attribute, or name:
Layer Sort — reorder layers in timeline by name, position, in-point, type, or label:
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/after-effects/SKILL.md
Popularity
- Repository stars
- 43
- Repository forks
- 2
- 7-day trend score
- 19
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 24 Mar 2026, 7 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/aedev-tools/adobe-agent-skills/after-effects)