Agent Skills: Create worktree + launch pi (default)

Launch one or more tasks in new git worktrees using worktrunk (wt).

Tasks: $ARGUMENTS

You are a dispatcher, not an implementer

FORBIDDEN: Do NOT read source files, edit code, or fix issues yourself. You only write prompt files and run wt switch commands.

If tasks reference earlier conversation (e.g., "do option 2"), include all relevant context in each prompt you write.

If tasks reference a markdown file (e.g., a plan or spec), re-read the file to ensure you have the latest version before writing prompts.

For each task:

1. Generate a short, descriptive branch name (2-4 words, kebab-case)
2. Write a detailed implementation prompt to a temp file
3. Run wt switch -c <branch-name> -x pi -- "$(cat <temp-file>)" to create worktree + launch agent (pi is default)

The prompt should:

• Include the full task description
• Use RELATIVE paths only (never absolute paths, since each worktree has its own root)
• Be specific about what the agent should accomplish

Skill delegation

If the user passes a skill reference (e.g., /some-skill), the prompt should instruct the agent to use that skill instead of writing out manual implementation steps.

Skills can have flags. If the user passes /some-skill --flag, pass the flag through to the skill invocation in the prompt.

Example prompt:

[Task description here]

Use the skill: /skill-name [flags if any] [task description]

Do NOT write detailed implementation steps when a skill is specified - the skill handles that.

Flags

--merge: When passed, add instruction to merge when done:

...
When complete, run: wt merge

--claude: Use Claude instead of pi:

wt switch -c <branch> -x claude -- "prompt here"

--opencode: Use OpenCode instead of pi:

wt switch -c <branch> -x opencode -- "prompt here"

--bg: Run in background tmux session (for handoffs):

tmux new-session -d -s <branch> "wt switch -c <branch> -x claude -- 'prompt'"

Workflow

Write ALL temp files first, THEN run all wt commands.

Step 1 - Write all prompt files (in parallel):

tmpfile=$(mktemp).md
cat > "$tmpfile" << 'EOF'
Implement feature X...
EOF
echo "$tmpfile"  # Note the path for step 2

Step 2 - After ALL files are written, run wt commands (in parallel):

wt switch -c feature-x -x pi -- "$(cat /tmp/tmp.abc123.md)"
wt switch -c feature-y -x pi -- "$(cat /tmp/tmp.def456.md)"

After creating the worktrees, inform the user which branches were created.

Remember: Your task is COMPLETE once worktrees are created. Do NOT implement anything yourself.

Quick Reference

# Create worktree + launch pi (default)
wt switch -c feature/auth -x pi -- "Implement OAuth flow"

# Create worktree + launch Claude
wt switch -c feature/auth -x claude -- "Implement OAuth flow"

# Create worktree + launch OpenCode
wt switch -c fix/bug-123 -x opencode -- "Fix the null pointer in auth.rs"

# Background session (handoff)
tmux new-session -d -s auth "wt switch -c feature/auth -x pi -- 'prompt'"

# Check status of all worktrees
wt list --full

# Merge completed work
wt merge