Claude-skill-registry forky
Fork Claude sessions to handle side tasks in parallel. Use when the user wants to spawn a parallel Claude session, delegate a task to a background agent, monitor forked sessions, or when they mention "forky", "fork", "spawn", or "parallel task".
git clone https://github.com/majiayu000/claude-skill-registry
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/forky" ~/.claude/skills/majiayu000-claude-skill-registry-forky && rm -rf "$T"
skills/data/forky/SKILL.mdForky - Parallel Claude Session Manager
Forky lets you spawn parallel Claude sessions to handle side tasks without interrupting your main workflow.
CRITICAL: Never Wait for Forks
Forks are fire-and-forget. After spawning, IMMEDIATELY continue with the next task. NEVER:
- Say "I'll wait for the fork to complete"
- Poll the fork status
- Defer work pending fork completion
Forks notify you when done. Just spawn and move on.
Spawning a Fork
ALWAYS run forky spawn as a background shell operation. This ensures you don't block waiting for the spawn to complete.
# Use run_in_background=true with the Bash tool ~/bin/forky spawn "your task description here"
The binary is at
~/bin/forky (NOT ~/.forky/bin/).
Model Selection
Always use Opus (the default). Do NOT use
-m haiku or -m sonnet without asking the user first.
When to Fork
Good:
- Code review while continuing development
- Research while implementing
- Test generation for code you just wrote
- Documentation tasks
- Any independent background work
Avoid:
- Tasks requiring immediate user input
- Modifying same files (use
if needed)--worktree - Very short tasks (overhead > benefit)
Monitoring
~/bin/forky serve # Start observability UI ~/bin/forky list forks # List all forks ~/bin/forky messages <id> # View fork messages
For Forked Agents
If you ARE a fork (check your callback instruction), call done when finished:
~/bin/forky done <your-fork-id> "summary of what was accomplished"