What problem does it solve?
Building scripts, coding agents, or test harnesses that consume Golem CLI output requires understanding its structured output contract: the --format flag, the $type discriminator, schema discovery, and stream framing. This Skill provides the parsing rules so tools interpret CLI output correctly instead of scraping human-readable text.
Core Features & Use Cases
- Structured Format Guidance: Explains how to request JSON, YAML, or TOON output via the global --format flag and why stdout carries the payload while stderr carries logs.
- Schema Discovery: Covers golem output-schema for listing $type values and fetching focused, pruned JSON Schemas for specific output types.
- Stream and Multi-Document Parsing: Details single, stream, and multi-document output modes, including TOON @toon/@end framing and line-delimited JSON streams.
- Use Case: You are writing a coding agent that invokes golem agent list and golem deploy. Use this Skill to branch on $type values, fetch focused schemas with golem output-schema --type, and parse streamed agent.oplog entries line by line.
Quick Start
Ask the AI to write a script that runs golem agent list with JSON output and parses the result using the $type field and output-schema metadata.