1
Define tool contracts
Specify validated inputs, outputs, errors, timeouts, and side-effect class.
Required evidence: Two read tools and one approval-gated write tool.
0/80 minimum characters
Advanced mini-project·90-120 minutes
Reliable Agent Harness Project
Design a bounded tool-using agent harness with schemas, budgets, approvals, deterministic mocks, and failure traces.
Scenario
An agent may read tickets and draft updates but must never send or mutate external systems without approval. Dependencies can time out or return malformed data.
You will demonstrate
- Specify tool permissions and schemas.
- Enforce call and retry budgets.
- Create reproducible failure tests and traces.
Project evidence
Show the work, not a checked box.
Each response is stored in this browser as you type. Include metrics, test output, or a decision rationale wherever the deliverable asks for it.
2
Enforce budgets
Limit total calls and retries and return an explicit terminal state.
Required evidence: Passing budget simulation tests.
0/80 minimum characters
3
Inject failures
Test timeout, malformed output, denied approval, and duplicate request.
Required evidence: Deterministic traces for every failure.
0/80 minimum characters
4
Define a release gate
Set task success, unsafe-action, and budget-exhaustion thresholds.
Required evidence: An automated go/no-go contract.
0/80 minimum characters
Runnable Python lab
Enforce a retry budget
Simulate tool outcomes and stop after success or a strict total-call budget.
Project defense
What is the safest default for a tool that sends an external message?
Rubric self-review
Rate the evidence, not your effort: 0 missing, 1 weak, 2 adequate, 3 strong. All criteria must be reviewed, but a low honest score does not get hidden.
Every side effect has an explicit trust boundary.
Budgets terminate loops deterministically.
Mocks make failures reproducible.
Traces support root-cause diagnosis without secrets.
Useful references
Project completion gate
Completion is controlled by stored evidence, deterministic tests, a decision defense, rubric review, and the artifact when required.
Evidence pendingCode pendingDefense pendingRubric pending
Optional cloud portfolio
Submit evidence across devices.
An account is required. Submit only when the local completion gate passes. AI review is advisory and separate from deterministic completion.