lazyllm-skill
LazyLLM framework for building multi-agent AI applications. Use when task mentioned LazyLLM or AI program for: (1) Flow orchestration - linear, branching, parallel, loop workflows for complex data pipelines, (2) Model fine-tuning and acceleration - finetuning LLMs with LLaMA-Factory/Alpaca-LoRA/Collie and acceleration with vLLM/LMDeploy/LightLLM. Includes comprehensive code examples for all components, (3) RAG systems - knowledge-based QA with document retrieval, vectorization, and generation, (4) Agent development - single/multi-agent systems with tools, memory, planning, and web interfaces.
graflow-workflow
Create Python workflow pipelines using Graflow with a structured plan-implement-review process. Use when building task graphs, parallel pipelines, LLM workflows, or any Graflow-based automation. Triggers on requests for "workflow", "pipeline", "task graph", "Graflow", or when user wants to build an automated data/AI pipeline.
open-dynamic-workflows
Give any coding agent dynamic-workflow power. Write a short JavaScript script in Claude Code's workflow dialect (export const meta + injected agent / parallel / pipeline / phase / log / args / budget globals) and run it against any coding-agent CLI (Codex, Claude Code, Gemini, Qwen, Kimi, or your own) with the `odw` command. Use when a task benefits from fanning out subtasks, multi-stage pipelines, adversarial verification, or loop-until-done discovery, rather than one in-context attempt.
Open Dynamic Workflow (odw)
odw is the **orchestration layer**. You write a small JavaScript workflow that calls `agent(prompt, opts)` nodes; odw schedules them (parallel / pipeline / budget / worktree / resume) and dispatches each node to **PandaCode**, which runs the actual AI (`codex`, `claude`, or `bamboo`) and owns accoun
ci-cd-integration
Design CI/CD pipelines for test execution. Covers GitHub Actions and GitLab CI pipeline templates, parallelism and sharding strategies, artifact management, flaky test quarantine, test result publishing, and quality gates. Includes copy-paste-ready workflow files for Playwright, Jest, and multi-stage pipelines. Use when: "CI/CD," "GitHub Actions," "pipeline," "test in CI," "GitLab CI," "continuous integration," "test automation pipeline." Related: playwright-automation, qa-metrics, self-healing-tests, coverage-analysis.
pipeline-subagents
Design and execute a custom multi-agent pipeline using Claude Code subagents directly. Spawns researcher, builder, etc. via the Agent tool with native parallel and background execution. Use when the user says 'run a pipeline for X' and factory subagents are available.
orchestrate-tickets
Coordinate claude-ui worker sessions over a batch of tickets through a staged pipeline: pull the backlog from the project's tracker, plan each ticket with an opus worker, implement it with a sonnet worker in its own git worktree, prove it with human-like Playwright UI testing and screenshot evidence, push and watch the CI build, open the pull request and get it reviewed by a fresh agent with the pr-review skill. Tracker, code host and CI are adapters (Jira, GitHub Issues, Bitbucket, GitHub, Jenkins, GitHub Actions); each repo supplies a ticket playbook. All findings land in local reports; the agent never comments on tickets or pull requests. Use when the user asks to orchestrate, run or work through a set of tickets or issues with parallel agents: "orchestrate these tickets", "work through the backlog", "chạy backlog", "spawn worker cho mấy ticket nà y". Needs a claude-ui session with orchestration on; the coordinator stays live to route worker questions.