Executions Commands
Inspect workflow, task, and agent executions with usage summaries.
Overview
The compozy executions
sub-commands surface real-time execution status, outputs, and the new LLM usage summaries. All commands support human-friendly table output as well as machine-readable JSON for automation.
Workflow Executions
Usage
compozy executions workflows get <workflow-exec-id> [flags]
Examples
# Show a workflow execution in the interactive table view
compozy executions workflows get 01JZMB7Z1B6KM0J6R2AVB59S4A
# Stream updates until completion
compozy executions workflows get 01JZMB7Z1B6KM0J6R2AVB59S4A --follow
Task Executions
Usage
compozy executions tasks get <task-exec-id> [flags]
Examples
# Fetch task execution details with usage metrics
compozy executions tasks get 01JZMB8ADP1T3QCW1SBBCX9KHF --format json --pretty
# Filter down to just the usage object
compozy executions tasks get 01JZMB8ADP1T3QCW1SBBCX9KHF \
--format json | jq '.usage'
Tip: The usage object matches the Execution Usage Schema, making it straightforward to push into billing or analytics pipelines.
Agent Executions
Usage
compozy executions agents get <agent-exec-id> [flags]
Examples
# Display an agent execution with the new usage summary field
compozy executions agents get 01JZMB8Z7F7T25N1M1SC1JW4S2 --format json | jq '.usage'
Filtering & Formats
All execution commands support the common flags:
Flag | Description |
---|---|
--format json | Emit JSON output suitable for automation. |
--pretty | Pretty-print JSON responses. |
--follow | Stream execution updates until completion. |
--quiet | Suppress progress logs in script mode. |
Related Resources
Workflow Commands
The workflow command group provides comprehensive workflow management capabilities including listing, inspecting, and executing workflows. These commands form the core of workflow orchestration in Compozy.
MCP Proxy Commands
Manage Model Context Protocol (MCP) servers and proxy operations with the `compozy mcp-proxy` binary. The MCP proxy server enables HTTP-based access to MCP servers, providing seamless integration between AI agents and external tools.