scitex-agent-container - Declarative AI Agent Lifecycle Management
scitex-agent-container is a declarative YAML-based framework for defining, managing, and orchestrating AI coding agent instances. Part of SciTeX.
Getting Started
Architecture & Reference
Ecosystem
API Reference
Key Features
YAML Definitions: Declarative agent configuration via YAML files
Lifecycle Management: Create, start, stop, and destroy agent instances
Rich Status:
status --jsonemits pane state, workspace files, hook-captured tool history, andlast_tool_at/last_mcp_tool_atfunctional-heartbeat shortcuts for dashboardsClaude Code Hook Integration:
hook-eventingests Claude CodePreToolUse/PostToolUse/UserPromptSubmit/Stopevents into a per-agent ring bufferZero Coupling: No knowledge of any downstream orchestrator; consumers (e.g. scitex-orochi) wrap
status --jsonto post to their own hubs
Quick Example
# Create and launch an agent from YAML definition
scitex-agent-container start my-agent.yaml
# List running agents
scitex-agent-container list
# Inspect live pane state / rich status
scitex-agent-container show-status my-agent --json
# Stop an agent
scitex-agent-container stop my-agent