bd-swarm-auto
Fully autonomous backlog completion. Runs waves until `bd ready` is empty. Self-resumable after /wipe. Use when you want hands-off parallel issue processing.
Installation
Details
Usage
After installing, this skill will be available to your AI coding assistant.
Verify installation:
skills listSkill Instructions
name: bd-swarm-auto
description: "Fully autonomous backlog completion. Runs waves until bd ready is empty. Self-resumable after /wipe. Use when you want hands-off parallel issue processing."
BD Swarm Auto - Autonomous Backlog Completion
YOU are the conductor. Execute this workflow autonomously. Do NOT ask the user for input.
Architecture: Fewer Terminals, More Subagents
Spawn 3-4 terminal workers max, each handling multiple issues via internal Task subagents.
BAD: 10 terminals x 1 issue each -> statusline chaos
GOOD: 3 terminals x 3-4 subagents each -> smooth execution
Quick Reference
| Step | Action | Reference |
|---|---|---|
| 1 | Get ready issues | bd ready --json |
| 2 | Calculate workers | Max 4, distribute issues |
| 3 | Create worktrees | Parallel, wait for all |
| 4 | Spawn workers | TabzChrome API |
| 5 | Send prompts | Multi-issue with subagent instructions |
| 6 | Poll status | Every 2 min, check context |
| 7 | Merge & cleanup | Kill sessions first |
| 8 | Visual QA | tabz-manager for UI waves |
| 9 | Next wave | Loop until empty |
Full details: references/wave-execution.md
EXECUTE NOW - Wave Loop
# Get ready issues
READY=$(bd ready --json | jq -r '.[].id')
[ -z "$READY" ] && echo "Backlog complete!" && exit 0
# Count and distribute to max 4 workers
ISSUES_COUNT=$(echo "$READY" | wc -l)
# 1-4: 1-2 workers, 5-8: 2-3 workers, 9+: 3-4 workers
# Create worktrees (parallel)
for ISSUE_ID in $READY; do
plugins/conductor/scripts/setup-worktree.sh "$ISSUE_ID" &
done
wait
# Spawn workers, send prompts, monitor
# ... see references/wave-execution.md
# Merge and cleanup
plugins/conductor/scripts/completion-pipeline.sh "$READY"
# Check for next wave
NEXT=$(bd ready --json | jq 'length')
[ "$NEXT" -gt 0 ] && echo "Starting next wave..." # LOOP
Key Rules
- NO USER INPUT - Fully autonomous, no AskUserQuestion
- MAX 4 TERMINALS - Never spawn more than 4 workers
- USE SUBAGENTS - Workers use Task subagents for parallelization
- YOU MUST POLL - Check issue status every 2 minutes
- LOOP UNTIL EMPTY - Keep running waves until
bd readyis empty - VISUAL QA - Spawn tabz-manager after UI waves
- MONITOR CONTEXT - At 70%+, trigger
/wipe:wipe
Worker Prompt Template
## Multi-Issue Worker Task
**MODE: AUTONOMOUS**
Do NOT use AskUserQuestion. Make reasonable defaults.
If blocked, close with 'needs-clarification' and create follow-up.
Issues to complete IN PARALLEL using subagents:
- ISSUE-001: Title
- ISSUE-002: Title
For EACH issue, spawn a Task subagent in ONE message.
Skills: /ui-styling:ui-styling, /frontend-design:frontend-design
Context Recovery
See: references/context-recovery.md
At 70% context, run /wipe:wipe with handoff:
## BD Swarm Auto In Progress
**Active Issues:** [list in_progress IDs]
**Action:** Run `/conductor:bd-swarm-auto` to continue
Auto vs Interactive
| Aspect | Auto | Interactive |
|---|---|---|
| Worker count | All ready | Ask user |
| Waves | Loop until empty | One wave |
| Questions | Make defaults | AskUserQuestion ok |
| Context | Auto /wipe | Manual |
Troubleshooting
| Problem | Solution |
|---|---|
| Worker not responding | tmux capture-pane -t SESSION -p -S -50 |
| Merge conflicts | Resolve manually, continue |
| Worker stuck | Nudge via tmux send-keys |
| Subagent failed | Worker retries or marks for review |
Reference Files
| File | Content |
|---|---|
references/wave-execution.md | Full 9-step execution details |
references/context-recovery.md | /wipe handoff procedure |
Execute this workflow NOW. Start with getting ready issues.
More by GGPrompts
View allSpawn and manage terminal tabs via TabzChrome REST API. Use when spawning workers, creating terminals programmatically, setting up worktrees for parallel work, or crafting prompts for Claude workers.
Patterns for building and debugging TabzChrome itself. Use when working on Terminal.tsx, xterm.js integration, WebSocket I/O, resize handling, or any TabzChrome extension/backend code.
Browser automation via 70 tabz MCP tools. Use when taking screenshots, filling forms, debugging network requests, testing responsive design, or using text-to-speech notifications.
This skill should be used when the user asks to "diagnose Claude Code", "fix MCP errors", "manage plugins", "configure MCP servers", "check /doctor errors", "debug plugin issues", "add MCP server", "remove plugin", "list plugins", "list MCP servers", "fix schema errors", or needs help with Claude Code CLI configuration and troubleshooting.