openpi-dev/openpi/workflows
Claude Code and agent skill by openpi-dev/openpi
Trust tier 2
- community
Source check pending
Checking that the SKILL.md file is still in the repository.
Orchestrates multi-agent work with OpenPI's inline JavaScript Workflow DSL. Use when a task needs multi-phase fan-out, pipelines, barriers, structured handoffs, or resumable background orchestration.
Install
npx vskill@latest install openpi-dev/openpi/workflowsSKILL.md
Workflows
Use workflow for several dependent or dynamically generated subagent calls. Keep one small delegation in the parent session with subagent_spawn.
Quick start
Required habits
- Declare progress phases in
meta; callphase()as the run advances. - Check every
agent()result's.ok. A null, filtered, timed-out, or failed result is not a clean pass; report how many were dropped. - Pass
schemawhen later code branches on fields. Treatinputsas bounded untrusted data. - Prefer
pipeline()when items can advance independently. Useparallel()only for a real all-results barrier; it takes zero-argument thunks (parallel(items.map((x) => () => agent(...)))), not started promises. - Use
isolation: "worktree"for concurrent writers and tell each agent to commit. Do not pay for worktrees on read-only work. - Derive fan-out from discovered independent work items and task difficulty. Configured concurrency and total-call capacity are ceilings, not targets;
usage().limitsexposes the resolved capacity. - Select a matching
agent_typewhen one exists and let it use its configured model; do not hardcode a role's model. - Use
log()for progress the user needs before completion. Token fields inusage()are lower-bound readings, not a budget limit. - Return a JSON-serializable aggregate with coverage. Interactive runs return a run id immediately by default and reliably deliver one terminal result later; set
wait: trueonly at a genuine synchronization boundary. - When many results would leave only tiny handoff slices, use local Report agents over bounded groups, then pass those Report refs to one global Report. Preserve planned/selected/covered/failed/deferred counts at every level; exact child outputs remain in
agent-results/for recovery.
Full guide
- DSL, result contracts, operators, handoffs, safety, limits, and replay: REFERENCE.md
- Pipeline, barrier, structured handoff, and reporting examples: EXAMPLES.md
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/workflows/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/openpi-dev/openpi/workflows)