Example ADL files for the ADL CLI

July 15, 2026 ยท View on GitHub

This directory contains example Agent Definition Language (ADL) files that demonstrate agent configurations across Go, Rust, and TypeScript, plus a range of deployment targets and optional features.

Every example here is part of the regression suite: task examples:test validates each manifest and task examples:generate scaffolds each one into test-output/.

Files

Go

  • go-agent.yaml - Go agent with services (database, notifications, reports), multiple tools, and an incident-response orchestration skill
  • go-agent-builtin-tools.yaml - Workspace assistant that exercises every reserved built-in tool (read, write, edit, bash, fetch) with sensible allow-lists for each
  • go-agent-ai.yaml - AI-powered Go agent with LLM-driven tools and the infer orchestrator
  • go-agent-artifacts-filesystem.yaml - Go agent with the filesystem-backed artifacts server enabled
  • go-agent-artifacts-minio.yaml - Go agent with the MinIO-backed artifacts server enabled
  • go-agent-telemetry.yaml - Go agent wired for OpenTelemetry - metrics server, OTLP traces, and per-tool-call spans
  • go-agent-documentation.yaml - Go agent that ships hand-authored documentation pages (spec.documentation.pages); the generator seeds stub docs/*.md files for the maintainers to fill in

Rust

  • rust-agent.yaml - Minimal Rust agent built with the Rust ADK, including an echo tool and a formatting skill
  • rust-agent-ai.yaml - AI-powered Rust agent with LLM-driven tool skills
  • rust-agent-redis.yaml - Rust agent with the Redis-backed task queue feature enabled

TypeScript

  • typescript-agent.yaml - Minimal TypeScript agent built with the TypeScript ADK
  • typescript-agent-tools.yaml - TypeScript agent with tools, services, and dependency injection
  • typescript-agent-ai.yaml - AI-powered TypeScript agent with LLM-driven tools
  • typescript-agent-telemetry.yaml - TypeScript agent wired for OpenTelemetry via the ADK's Node SDK provider

Deployment targets

  • cloudrun-agent.yaml - Go agent configured for Google Cloud Run using Google Container Registry (spec.deployment.type: cloudrun)
  • cloudrun-ghcr-agent.yaml - Go agent configured for Google Cloud Run using GitHub Container Registry
  • cloudflare-agent.yaml - TypeScript agent configured for Cloudflare Workers (spec.deployment.type: cloudflare)
  • kubernetes-agent.yaml - Go agent configured for Kubernetes deployment via the inference-gateway operator (spec.deployment.type: kubernetes)
  • vercel-agent.yaml - TypeScript agent configured for Vercel deployment from source (spec.deployment.type: vercel), exercising the edge runtime, regions, function limits, and env vars

Usage

Validate an example manifest:

adl validate examples/go-agent.yaml

Generate a project from an example:

# Go agent
adl generate --file examples/go-agent.yaml --output ./test-go-agent

# Rust agent
adl generate --file examples/rust-agent.yaml --output ./test-rust-agent

# TypeScript agent
adl generate --file examples/typescript-agent.yaml --output ./test-typescript-agent

Deployment targets are selected by the manifest's spec.deployment.type; the matching --deployment flag is optional and simply asserts the expected target:

# Cloud Run
adl generate --file examples/cloudrun-agent.yaml --output ./test-cloudrun-agent --deployment cloudrun

# Kubernetes
adl generate --file examples/kubernetes-agent.yaml --output ./test-kubernetes-agent --deployment kubernetes

Layer on CI and/or CD pipelines with --ci and --cd:

adl generate --file examples/go-agent.yaml --output ./enterprise-agent --ci --cd

Validate or generate every example at once via the Taskfile:

task examples:test      # validate all manifests
task examples:generate  # scaffold every example into test-output/