beam-master_skill

Overview

This skill is a shared resource library that centralizes setup, authentication, API references, error handling, and utility scripts for Beam AI integrations. It is not intended to be invoked directly; instead it supports specific Beam integration skills by providing a single source of truth. The library reduces duplication and simplifies maintenance across multiple Beam tools and scripts.

How this skill works

The skill collects reference documentation (setup, API patterns, authentication, error handling) and a suite of CLI scripts for common operations like token exchange, agent management, graph/node operations, task lifecycle, and HITL actions. Integration skills import or reference these resources to run pre-flight checks, perform authenticated requests, create and monitor tasks, and debug failures. Scripts provide consistent CLI interfaces and exit codes for automation and debugging.

When to use it

• When you need a single authoritative guide for Beam API patterns, authentication, and troubleshooting.
• Before running any Beam operation to validate configuration and credentials.
• When building or updating Beam integration skills to avoid duplicating setup and error handling logic.
• When debugging authentication, token lifecycle, or task execution issues across Beam tools.
• When you want CLI utilities to list agents, inspect graphs, run or retry tasks, and handle HITL flows.

Best practices

• Run the configuration check script first to confirm API key and workspace ID are valid before other operations.
• Use the interactive setup wizard to create or refresh credentials and persist them in .env for CI and local use.
• Prefer token-exchange and refresh scripts when debugging auth issues; rely on automated token handling in production clients.
• Use cursor-based iteration scripts for large exports instead of paginated list endpoints to avoid partial results.
• Log and surface structured JSON output (--json) from scripts for reliable parsing by automation or monitoring tools.

Example use cases

• Validate workspace access and troubleshoot auth errors before launching agent tasks for a user.
• List agents and fetch an agent's graph to identify node IDs for targeted testing or updates.
• Test an individual graph node with custom input to diagnose incorrect outputs or failing integrations.
• Create, monitor, and stream updates for an agent task, including approving or rejecting HITL requests.
• Retry failed tasks after resolving transient errors or exporting large task sets for analysis using iterate scripts.