What problem does it solve?
Interacting with a running Friday daemon for operational tasks like managing workspaces, triggering agent workflows, and inspecting session outcomes often requires memorizing complex CLI flags or manually crafting raw HTTP requests. This Skill serves as a single, authoritative reference for both the atlas CLI and the daemon's HTTP API, eliminating guesswork and reducing errors when working with Friday's runtime.
Core Features & Use Cases
- Dual Surface Coverage: Provides complete reference material for both the human-friendly atlas CLI (for common inspection, triggering, and monitoring tasks) and the raw HTTP API (for advanced operations like partial workspace config updates, SSE streaming, and MCP registry mutations not exposed via the CLI).
- Workflow Automation Support: Includes end-to-end recipes for common automation patterns, including creating workspaces from YAML configs, firing and monitoring signals, publishing and scoping skills, registering custom Python agents, and implementing the autopilot self-modification flywheel.
- Safety Guardrails & Troubleshooting: Outlines clear guardrails for destructive operations, common error codes and their resolutions (e.g., 409 conflicts from active sessions, 422 payload schema errors), and guidance on when to use CLI vs HTTP for different tasks.
Quick Start
Use this skill to verify your local Friday daemon is running, list all registered workspaces, trigger a test signal on a workspace of your choice, and stream the live session events to confirm the associated agent workflow executes successfully.