flowing - Execute multi-step Python DAG workflows with control flow
Encodes multi-step procedures as Python task DAGs with enforced ordering, branching, retries, input validation, and checkpoint resuming.
Tags
Updated: 2026-09-17Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Execute Python task DAG workflows
- Enforce task dependencies and ordering
- Branch tasks using predicate conditions
- Validate inputs before task execution
- Retry tasks until conditions pass
- Run independent tasks in parallel
- Execute detached non-blocking side-effects
- Resume execution from durable journal
Inputs
- Python task definitions
- Task dependency graph specifications
- Journal file path for checkpoints
Outputs
- Task output values and side-effects
- JSONL checkpoint journal file
Requirements
- Python runtime
- flowing.py script in scripts directory