Use at conversation start when user mentions: parallel, multiple tasks/features, worktrees, or lists 2+ independent work items - determines which polydev skill to use
/plugin marketplace add shikihane/polydev/plugin install polydev@polydev-marketplaceThis skill inherits all available tools. When active, it can use any tool Claude has access to.
This is NOT optional. If there's even a 10% chance polydev applies, CHECK IT. </CRITICAL>
All scripts MUST be called via $POLYDEV_SCRIPTS variable. NEVER use ./scripts/
POLYDEV_SCRIPTS="/path/to/polydev/plugins/polydev/scripts"
"$POLYDEV_SCRIPTS/spawn-session.sh" <workspace> <branch> <worktree-path> <plan-file>
User message received
|
Contains parallel/multiple keywords?
| YES
Is it complex/unclear? --YES--> Run /polydev-brainstorm
| NO
Ready to execute? --YES--> Use polydev:polydev skill
|
Need detailed plans? --YES--> Use polydev:writing-plans skill
|
Running background command? --YES--> Use polydev:terminal-task-runner skill
| Skill | When to Use | Who Uses |
|---|---|---|
/polydev-brainstorm | Complex/unclear requirements, need to decompose | Main Agent |
polydev:polydev | Ready to execute parallel tasks | Main Agent |
polydev:writing-plans | Need detailed implementation plans | Main Agent |
polydev:terminal-task-runner | Long-running commands (builds, tests, servers, SSH) | Main Agent |
polydev:worktree-executor | Execute in isolated worktree | Sub-Agent only |
polydev:agent-investigator | Read-only research tasks | Sub-Agent only |
| Scenario | Script | Parameters |
|---|---|---|
| Create worktree + Claude | spawn-session.sh | <workspace> <branch> <worktree-path> <plan-file> |
| Monitor status | poll.sh | <worktrees-dir> <timeout> |
| Restore crashed session | restore-session.sh | <worktree-path> [--force] |
| Send to worktree | wo-send-command.sh | <worktree-path> "<cmd>" |
| Send to any session | send-to-session.sh | <session_id> "<cmd>" |
| Read screen output | capture-screen.sh | --session <wo:id> --lines N |
| List sessions | list-sessions.sh | [workspace] |
| Close session | close-session.sh | <session_id> |
| Start background command | run-background.sh | <name> "<cmd>" |
| Analyze output | analyze-output.sh | <session_id> --lines N |
| Wait for pattern | wait-for-pattern.sh | <session_id> --success "<pattern>" |
| Start investigation agent | spawn-agent.sh | <name> --prompt "<task>" --report <path> |
If you catch yourself thinking:
| Thought | Reality |
|---|---|
| "I'll just do these sequentially" | If independent, parallelize. Check polydev. |
| "This is simple enough to do directly" | 2+ tasks = potential parallelism. Check. |
| "I don't need the overhead" | Polydev saves time on multi-task work. |
| "Let me explore first" | Run /polydev-brainstorm to explore properly. |
| "I'll parallelize later" | Parallelize NOW if tasks are independent. |
All of these mean: Check polydev skills first.
User says "implement X, Y, and Z":
User says "run this build/test":
User says "SSH to server and run commands":
run-background.sh to start SSHsend-to-session.sh to send subsequent commandscapture-screen.sh to read outputUser says "research X":
Sub-agents MUST use model: "sonnet" unless user explicitly requests otherwise.
// Correct
Task({ prompt: "...", subagent_type: "general-purpose", model: "sonnet" })
// Wrong - will bankrupt user
Task({ prompt: "...", subagent_type: "general-purpose" })
This skill is the entry point. After determining which skill to use:
using-polydev (this skill)
|
/polydev-brainstorm (if complex)
|
polydev:writing-plans (if need plans)
|
polydev:polydev (execution)
|
polydev:worktree-executor (per-branch, sub-agent)
This skill should be used when the user asks to "create a slash command", "add a command", "write a custom command", "define command arguments", "use command frontmatter", "organize commands", "create command with file references", "interactive command", "use AskUserQuestion in command", or needs guidance on slash command structure, YAML frontmatter fields, dynamic arguments, bash execution in commands, user interaction patterns, or command development best practices for Claude Code.
This skill should be used when the user asks to "create an agent", "add an agent", "write a subagent", "agent frontmatter", "when to use description", "agent examples", "agent tools", "agent colors", "autonomous agent", or needs guidance on agent structure, system prompts, triggering conditions, or agent development best practices for Claude Code plugins.
This skill should be used when the user asks to "create a hook", "add a PreToolUse/PostToolUse/Stop hook", "validate tool use", "implement prompt-based hooks", "use ${CLAUDE_PLUGIN_ROOT}", "set up event-driven automation", "block dangerous commands", or mentions hook events (PreToolUse, PostToolUse, Stop, SubagentStop, SessionStart, SessionEnd, UserPromptSubmit, PreCompact, Notification). Provides comprehensive guidance for creating and implementing Claude Code plugin hooks with focus on advanced prompt-based hooks API.