Your First Harness Proof Packet
Turn a harness run into a portfolio artifact with architecture, receipts, evals, and a short engineering narrative.
A proof packet is the artifact that makes your agentic engineering legible. It should help a reviewer understand what you built, how it ran, what it was allowed to do, what evidence was saved, and what you learned from the result.
A simple proof packet includes:
- A one-page architecture diagram.
- The agent goal and pact.
- The tool registry and permission notes.
- A run receipt with tool calls and key outputs.
- A token budget estimate and actual usage.
- An evaluation result or checklist.
- A short narrative explaining one failure and one improvement.
The narrative is important. Early-career candidates often show finished outputs but skip the engineering judgment. A strong proof packet says: here was the risk, here was the boundary, here was the evidence, here is what changed after verification.
This is the bridge into the paid certification. The free prep course teaches the shape of the artifact. The certification adds reviewed harness runs, a private async critique loop, and a final portfolio packet designed to hold up in internships, interviews, or internal career conversations.
Course complete
Agentic Harness Engineering Prep
Go deeper with certification
AI Agentic Harness Engineering Certification - async online, $1,000
Enroll nowNew courses drop every few weeks
Get notified when new content goes live — no spam, unsubscribe any time.
Start building trusted agents
Register an agent, define behavioral pacts, and earn a verifiable TrustMark score.