Practical notes: Testing the OpenAI Agents API with a Hosted Sandbox

The following notes reconstruct a practical path around “Testing the OpenAI Agents API with a Hosted Sandbox”. Emphasis stays on contracts, checks, and drop-in code placeholders rather than motivational framing. When working through the Overview stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

01 How the Agents API fits together

The 01 How the Agents stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

02 Why we chose a hosted sandbox

The 02 Why we chose stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

Hosted settings that matter here

The Hosted settings that matter stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts. The Hosted settings that matter stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

03 Inside our CSV runner

For the 03 Inside our CSV stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

The actual seed-42 input

For the The actual seed-42 input stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

amount,note
 $981.96 ,sale
" $1,838.10 ",sale
 $42.78 ,sale
-$100.12,refund
N/A,voided
 $155.92 ,sale

{
  "total_cents": 291864,
  "valid_rows": 5,
  "invalid_rows": 1
}

The instructions we send

For the The instructions we send stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness. For the The instructions we send stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

What we check and save

When working through the What we check and stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

04 What the five recorded trials show

When working through the 04 What the five stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

What changed in our runner

When working through the What changed in our stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node. When working through the What changed in our stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

05 How our runner handles failures

The 05 How our runner stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

06 When to bring your own compute

The 06 When to bring stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

07 Run our experiment

The 07 Run our experiment stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts. The 07 Run our experiment stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

cd openai-agents-api
python -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements.txt
export OPENAI_API_KEY="your-key"
python sandbox_check.py --seed 42       # Start with one trial
# python sandbox_check.py 5 --seed 42   # Optional: five trials

trial-001/
  data.csv           # Exact input
  expected.json      # Local reference
  request.json       # Submitted configuration
  session.json       # Latest retrieved session
  turns.json         # Turn status, timestamps, usage
  items.json         # Saved messages and tool calls
  artifacts.json     # Published file metadata
  summary.json       # Downloaded result
  calculate.py       # Downloaded agent script
  result.json        # Checks, timing, errors, cleanup

python -m unittest discover -s . -p test_sandbox_check.py -v

Operational checklist

For the Operational checklist stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state.

Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments.

Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

Write a short runbook: how to rotate keys, how to drain the queue, how to roll back the last ingest.

Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

Before promoting the stack, freeze versions, capture a golden transcript for the critical path, and confirm rollback steps. Shared environments need rate limits, tenancy checks, and a clear owner for secret rotation. Prefer boring reliability over clever one-off demos.

Batch note for f317619e24fa: keep provider keys out of the repo, set a per-session token ceiling, and store transcripts next to the eval fixtures so later model swaps stay comparable.

When working through the hardening note 0 stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish.

Hardening detail 0/838: measure wall time, error class, and token spend for this note, then decide whether to keep the change based on a fixed question set rather than anecdote.

The hardening note 1 stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

Hardening detail 1/838: measure wall time, error class, and token spend for this note, then decide whether to keep the change based on a fixed question set rather than anecdote.