Orchestrate multi-agent LLM workflows with declarative YAML DAGs and centralized coordination.
npx skills add https://github.com/zendev-sh/zenflow --skill zenflow
Please help me install this Agent Skill. Skill: zenflow Source: https://github.com/zendev-sh/zenflow/tree/main Command: npx skills add https://github.com/zendev-sh/zenflow --skill zenflow
This Skill solves the complexity of managing multi-agent LLM pipelines by providing a declarative, race-safe, and observable framework for orchestrating agent workflows.
Use the zenflow skill to execute the workflow defined in review.yaml using the gemini-3-pro-preview model.
High-intent search queries and answers about installing and using this skill.
You can orchestrate multi-agent LLM workflows declaratively by defining complex DAGs with parallel fan-out, loops, and conditions in YAML, eliminating the need for imperative control flow code.
Race-safe message delivery between agents is handled through a hub-and-spoke messaging architecture using a centralized coordinator, which prevents silent message loss during distributed workflow execution.
Yes, you can define branching, loops, and conditions directly within a declarative YAML specification to manage complex automation tasks across distributed agent steps with persistent state management.
zenflow supports native MCP tool integration and outputs structured NDJSON events, providing observability for multi-agent pipelines while maintaining persistent state across distributed steps.
The best way is to define the lifecycle in a single YAML specification where a planner agent defines tasks, a coder implements features, and a reviewer validates changes, all coordinated by a centralized executor.
You should avoid declarative YAML orchestration if your multi-agent tasks do not require complex DAGs, branching, loops, or persistent state management across distributed steps and can be handled by simple sequential calls.
ThrookieAutomate Claude Code loop design and execution for multi-agent tasks.
balkhaevAnalyze NDJSON application logs to identify errors and performance issues.
chemistrywow31Analyzes agent graph topology for safe handoffs and parallel work in Codex multi-agent systems.
RachasumanthDefine and execute multi-agent workflows with a declarative programming language.
