Use lessons 01–05 or equivalent skills before this exercise. You can revisit them from the lesson list.
A computer and Python 3.10+. The supplied standard-library lab needs no API key, network or purchase.
Download the runnable offline builder lab →Start with the idea
Deliver an offline workflow with honest execution evidence.
Record the setup, commands, results and remaining gaps. Inspect the actual output. Keep offline control tests separate from claims about live models or real use.
Release note: A record of the version, tested behavior, known gaps and recovery plan.
See the example and check your understanding
A made-up teaching example
The guest public lookup produces P at 10:00. The restricted lookup is denied. The lab’s control tests pass, but persistence and live model behavior were not tested.
Ask yourself
Does passing this lab prove that an autonomous service is production-ready?
Show the explanation
No. Report only what was tested. A denied request must produce no output. Label work that was not run.
Start with these checks
- Inspect actual source IDs and values.
- Check denial produces no output.
- Separate tested behavior from proposed work.
Use the different practice case below and check every requirement in its full instructions.
Full learning objectives
- Produce an inspectable artifact.
- Test unsupported and denied paths.
- Explain configuration and deployment limits.
Captions appear here during playback.
Capstone: deliver a bounded offline workflow
Deliver an offline workflow with honest execution evidence.
Video playback is unavailable. Read the descriptive transcript below, go directly to practice, or open the video file.
Source-based teaching example · Fictional data · Reviewed 2026-10-03. Microsoft Mark · locally synthesized. On a small screen, use fullscreen for diagram details or read the transcript.
Read the descriptive transcript
0:00 · What you will learn
Deliver an offline workflow with honest execution evidence.
What appears on screen
Deliver an offline workflow with honest execution evidence.
0:06 · The idea
Record the setup, commands, results and remaining gaps. Inspect the actual output. Keep offline control tests separate from claims about live models or real use.
What appears on screen
Record the setup, commands, results and remaining gaps. Inspect the actual output. Keep offline control tests separate from claims about live models or real use.
0:21 · One useful term
Release note. A record of the version, tested behavior, known gaps and recovery plan.
What appears on screen
A record of the version, tested behavior, known gaps and recovery plan.
0:31 · A worked example
The guest public lookup produces P at 10:00. The restricted lookup is denied. The lab’s control tests pass, but persistence and live model behavior were not tested.
What appears on screen
The guest public lookup produces P at 10:00. The restricted lookup is denied. The lab’s control tests pass, but persistence and live model behavior were not tested.
0:45 · Your turn to think
Does passing this lab prove that an autonomous service is production-ready? Pause and decide before continuing.
What appears on screen
Does passing this lab prove that an autonomous service is production-ready? Pause the video if you need more time.
1:03 · Why this answer works · 1
No. Report only what was tested. A denied request must produce no output.
What appears on screen
No. Report only what was tested. A denied request must produce no output.
1:12 · Why this answer works · 2
Label work that was not run.
What appears on screen
Label work that was not run.
1:17 · Check 1
Inspect actual source IDs and values.
What appears on screen
Inspect actual source IDs and values.
1:23 · Check 2
Check denial produces no output.
What appears on screen
Check denial produces no output.
1:28 · Check 3
Separate tested behavior from proposed work.
What appears on screen
Separate tested behavior from proposed work.
1:34 · Now try a different case
Open the lesson and try its different practice case. Use every check in the full task. Watching alone does not mark the lesson complete.
What appears on screen
Open the lesson Try the new case Check and revise The full task and its requirements are in the linked lesson.
Watching explains the example. Complete the lesson practice to check what you can do. Explore all visual lessons →
The method
Integrate the parser, lookup, answer validator and operation cache into run(caller, request, records, operation_id, cache). Reauthorize every request before cache reuse. Bind an operation ID to the caller, exact request and current source time/scope: a changed intent under that ID returns key_conflict; use a new ID for a new operation. Only an allowed successful lookup may generate an artifact. Run the integrated repeat, source-change and revoked-permission checks, and keep your final claim consistent with actual output. The cache remains an in-process teaching fixture. Use the reference only after attempting implementation.
Prepare a release note with command output, artifact inspection, negative-case results and remaining limitations. The workflow is intentionally deterministic and offline. It checks application controls, not live model quality, provider compatibility, internet access or production security. An optional future model component needs a documented adapter and fresh evaluation before making additional claims.
Choose your model, harness and offline route
Text-only assistant
Use supplied text. Ask for a plan or draft; do not claim browsing, code execution or file access that was unavailable.
Tool-enabled application or harness
Record which tools actually run, allowed files and actions, approval boundaries and observable results. Tools are optional unless the task needs them.
Local or cloud deployment
Inspect inference, retrieval, connectors, logs and storage separately. Use only authorized environments for real data; training fixtures are fictional.
Higher-capability or frontier model
Evaluate the exact task and configuration. Capability does not supply missing evidence or grant permission. No provider ranking or guarantee is implied.
Offline alternative
Use the supplied inputs and inspect the worked example manually. Builder exercises use local deterministic Python. No paid account or live model call is required.
Record the exact product/model label when available, tools actually used, inputs, permitted actions, observed output and checks. Live accounts or APIs may have external costs; they are optional and are not used by this page.
Read the general workflow guide →Worked example
Supplied input
Worked release note structure: environment; intended behavior; command executed; case results; artifact inspected; unresolved cases; rollback. Explicit label: offline fixture, not provider benchmark.
A checked result
“The guest public lookup produced P at 10:00; guest restricted lookup was denied. Six control groups passed in this isolated lab. Persistence and live model behavior were not tested.”
Why this works
A narrow truthful release note is stronger evidence than claiming an autonomous agent is production-ready.
Your independent capstone
Independent capstone: finish run in starter.py; add a public record N with time 14:15 and an independent test. Run python check.py, inspect the returned artifact and record one failed case you repaired.
Download this task and checklist →Interactive practice / Fictional fixture
Inspect a bounded pipeline artifact
Run a browser simulation against public N, restricted R or an absent ID. This is a deterministic teaching pipeline, not a model run or production deployment.
Everything runs in this browser. No AI model is called and no work is uploaded. Bounded checks cover specific facts and rules. Your writing is not graded: you review it yourself. Practice never marks a lesson complete automatically.
Record observed browser gates and artifact separately from actual Python tests. Name at least one untested production property.
Try the task, run the checks, then revise. Checks are teaching feedback, not an instructor grade.
Observed browser simulation
This trace comes from the bounded local checker. It is not a Python run, model result or production tool execution.
Keep or restore your work
Drafts are not saved automatically. Save only fictional practice data. A browser draft may be lost if storage is cleared; download a copy to keep it.
Your steps
- Integrate only the supplied offline operations.
- Verify public, denied, absent and malformed paths.
- Add and execute the new N case.
- Write a release note including your diff, commands, artifact and limitations.
Check your work
- Correct public artifact includes the source record and time.
- Denied paths return no artifact.
- Actual execution is distinguished from proposed work.
- No external calls, payments or production claims are required.
After your attempt: review notes and a common mistake
The acceptance package needs both positive and negative paths plus an inspected artifact. The supplied checker is a teaching aid, not independent certification; preserve your added case and limitations.
Different wording or a different justified decision can pass. Check your work against every requirement; these notes are not an automated grade.
The lab’s reference.py is a worked solution. Run python check.py --reference only to inspect feedback after your attempt; passing the reference does not complete your work.
Explain your approach
How would your evaluation change when replacing the deterministic answer with a real model?
Keep your artifact, one correction or decision, its evidence and one remaining limitation. Nothing on this page sends or uploads your work.
Check your understanding
Choose an answer to see feedback. This is a knowledge check, not instructor assessment.
Review your independent practice
Complete the supplied task before recording progress. Use the acceptance criteria and reveal the review notes after your attempt. Keep your artifact yourself; nothing is uploaded.
Not yet self-reviewed. Completion stays in this browser and does not issue a badge.
Optional deeper study
Explore the related general-course lesson → This opens the existing canonical lesson; it is optional depth, not another copy of this focused task.
Primary references
Revision 2026-10-03. Fictional examples, not observed provider results. Knowledge checks and completion records are self-study aids.