add-gif-gen

name	add-gif-gen
description	Add the gif-gen container skill (crop, convert, slideshow, sprite sheet). Free — uses ffmpeg locally inside the container. No API keys.

Add GIF Generation (free modes)

Installs the gif-gen container skill (one Bash script, four modes — all ffmpeg/ffprobe local). No OneCLI secret. No cost.

This is the free subset of the v1 gif-gen pack. The --animate mode (fal.ai Kling, $0.42 per call) is intentionally not ported.

Phase 1: Pre-flight

Already applied?

test -f container/skills/gif-gen/generate-gif && echo INSTALLED || echo NOT_INSTALLED

If INSTALLED, skip to Phase 4 (Verify).

Phase 2: Install the container skill

rsync -a .claude/skills/add-gif-gen/container-skills/ container/skills/
chmod +x container/skills/gif-gen/generate-gif
head -5 container/skills/gif-gen/SKILL.md

Phase 3: Ensure ffmpeg in the container image

The script needs ffmpeg and ffprobe (same package). Two paths:

3.1 Per-group via install_packages (recommended)

If the agent group already has ffmpeg (e.g. you installed /add-voice first, which also needs it), skip. Otherwise, ask the agent to install it:

Send the agent: "add ffmpeg to your container packages". The agent calls mcp__nanoclaw__install_packages({apt: ["ffmpeg"]}). Approve in the host UI. The container rebuilds automatically.

3.2 Global Dockerfile (alternative)

If you want ffmpeg baked into the base image for every group:

grep -q 'ffmpeg' container/Dockerfile && echo "PRESENT in Dockerfile" || \
  echo "Edit container/Dockerfile apt-get install line to add 'ffmpeg', then run ./container/build.sh"

After rebuilding, all groups get ffmpeg without per-group install.

Phase 4: Sync skills to existing groups

for session_dir in data/v2-sessions/ag-*; do
  if [ -d "$session_dir/.claude-shared/skills" ]; then
    rsync -a container/skills/ "$session_dir/.claude-shared/skills/"
    echo "Synced skills to: $session_dir"
  fi
done

Phase 5: Restart running containers

docker ps --format "{{.ID}} {{.Names}}" | grep nanoclaw-v2 | awk '{print $1}' | xargs -r docker stop

Phase 6: Verify

Send to any agent (with ffmpeg available):

"toma esta imagen y conviértela a gif" (with an image attached) → expect Bash(generate-gif --convert ...).

Or for a slideshow with multiple images:

"haz un gif con estas tres imágenes" → expect Bash(generate-gif img1 img2 img3).

Troubleshooting

generate-gif: command not found: skill not synced to the agent group. Re-run Phase 4.

ffmpeg: command not found: ffmpeg is not in the agent's container. Run /add-voice (which installs it) or follow Phase 3 above.

Output GIF is too small: edit scale=480:-1 in the script to a larger value (e.g. scale=720:-1). Bigger GIFs = bigger files = slower delivery.

Sprite-sheet output is mis-aligned: image_width / cols or image_height / rows produces a fractional frame size. Re-check the grid by looking at the source image with vision.