Example: an Agent that builds another Agent (local, on an AMD GPU via Ollama)

This example is the "Harness for Building Agents" pillar in runnable code. Using only the PenguinHarness SDK, it:

builds a brand-new agent (commit-helper) by driving default_agent with the agent-creation skill from a plain-language requirement, then
runs that freshly-created agent to prove the generated AGENTS.md actually shapes its behavior (it writes a Conventional Commits message).

Everything runs on a local open-weight model — qwen3.6:35b served by Ollama — so no cloud API and no data leaving the machine. Ollama's ROCm backend runs this natively on AMD GPUs (from Radeon PRO workstation cards up to Instinct accelerators).

1. Serve the model locally with Ollama

# Ollama detects an AMD GPU (ROCm) automatically; pin a specific card if you like:
export HIP_VISIBLE_DEVICES=0
ollama serve &          # if not already running as a service
ollama pull qwen3.6:35b

2. Point PenguinHarness at it (once)

penguin config model add \
  --model-id qwen3.6:35b \
  --provider custom --client-type openai \
  --base-url http://localhost:11434/v1 \
  --api-key ollama --set-default

This writes the model into ~/.penguin/data/default_project/.project_config.toml. The example uses the project's default model, so no model id is hard-coded in the script.

3. Run the example

From the repo root (build the workspace first so @prismshadow/penguin-core resolves to its dist/):

pnpm install
pnpm build
pnpm --dir examples/build-agent-with-agent start
# or directly:  npx tsx examples/build-agent-with-agent/build-agent.ts

What you should see

Phase 1 — default_agent scaffolds agents/commit-helper/ under the project: its directory layout, a copied system_config.yaml (with name + description), and an AGENTS.md encoding the Conventional Commits rules.

Phase 2 — the new commit-helper agent, following only that generated AGENTS.md, produces something like:

fix(payment): add retry-with-backoff for transient gateway 503 errors

Transient 503 responses from the payment gateway were causing checkout
failures during peak traffic. Retry with exponential backoff gives the
gateway time to recover, preventing spurious user-facing errors.

Notes

Output quality tracks the model. qwen3.6:35b handles this task well; much smaller models may not follow the tool-calling protocol reliably.
A capable model occasionally introduces a mechanical slip when re-serializing the base config (e.g. an invalid YAML escape) — every file is plain text and every run is traced, so such slips are quick to spot and fix. This is the realistic shape of "agents building agents": the requirement→AGENTS.md heavy lifting is automated; a human reviews the mechanical edges.
Re-running Phase 1 will update the existing commit-helper agent in place.

3.0 KiB Raw Blame History