CCAR-F / Free worksheet and question practice

Will the next phase receive the context it needs?

Write a handoff that keeps separate issues and their evidence intact. Then revise it when one source changes, carry the updated context into the next phase, and explain what your finite local checks cannot establish.

Use Python 3.10 or newer and its standard library. No package installation, account, key or network service is needed after download. You edit a JSON worksheet and assess its prose with a human rubric. Your files and results stay on your computer; this page collects no submissions or completion history.

Download the handoff worksheet (ZIP)

Open the worksheet locally

  1. Check python3 --version (Windows: py -3 --version).
  2. Extract the worksheet into a new empty folder and open a terminal in that folder. Read README.md, CONTRACT.md and WORKSHEET.md.
  3. Read the supplied source fixture. Copy handoff.json to MY_HANDOFF.json and edit the copy. Keep the fixture and checker unchanged. The starter is intentionally incomplete.
  4. Follow the worksheet’s digest and trace steps, then run the contract check below. Windows uses py -3 in place of python3.
python3 check_contract.py --digests MY_HANDOFF.json
python3 check_contract.py MY_HANDOFF.json
python3 check_contract.py --self-test

Digest output helps bind the local artifacts; it does not validate the packet. The self-test exercises the checker’s finite acceptance and rejection cases; it does not grade your handoff. Use a disposable folder separate from customer or production work.

Write, carry forward, then revise

  1. Keep issues separate. Record the facts and source identity for each issue. Put the findings up front, with headed details that a person can inspect.
  2. Supply the next phase. Copy the first phase’s required evidence and facts into the second phase’s initial context. Record the explicit save → prepare → load → simulated dispatch sequence bound to the artifact digests.
  3. Handle the changed case. Source A advances while source B is preserved. Update the affected facts and context without silently replacing the other issue’s evidence.
  4. Explain the boundary. Use the human rubric to assess what the handoff communicates, what remains unresolved and why a passing structural check is insufficient.

A person assesses the reasoning

Open RUBRIC.md and review your original and changed rounds. Assess whether the findings are useful, issues remain distinct, evidence supports the claims, and the next phase can understand the changes and limitations. Write your reasons in a separate local MY_REVIEW.md rather than treating the checker’s exit code as a reasoning score.

The checker validates a finite shape, source identities, artifact digests, context presence and simulated ordering. It does not judge prose adequacy, entailment, real dispatch, durable storage or a real agent’s use of context.

Compare after your attempt

Keep your worksheet. Extract the separate worked answer into another empty folder, open a terminal there, and check its supplied handoff.json as shown below. Compare its walkthrough and rubric reasoning with your own.

Reveal the worked answer download

The worked answer is one example for these supplied fixtures. Its passing contract check does not establish that its prose is uniquely correct or generally effective.

python3 check_contract.py handoff.json
Download the separate worked answer (ZIP)

Try three explained practice questions

Apply the worksheet ideas to three original synthetic decisions: separate issue facts, organize findings, and supply the next phase’s initial context. Each question has one answer under its stated contract. Use the explanations to revisit your handoff and human rubric.

This is free practice with independently reviewed explanations. It covers only part of these skills and does not measure mastery, exam readiness or pass probability. These are not official exam questions. Your selections stay in this page’s memory and disappear when you reload; no answers or completion history are saved or submitted.

Practice question 1

A synthetic support investigation covers two issues. ISSUE-A concerns order SYN-1041, GBP 48.00, a carrier exception and an unknown parcel location. ISSUE-B concerns order SYN-2084, GBP 12.50, a requested invoice correction and an unknown approved replacement name. Each observation has its own source revision and observation time. A compacted narrative says only “delivery and invoice problems need sorting.” Which context design best preserves the facts needed to continue both issues independently?

Select one answer, then check it to read the explanation. Enable JavaScript to select and check answers.

Practice question 2

An aggregation step must prepare the synthetic investigation for a reviewer. The team requires a short key-findings summary at the beginning, followed by explicit per-issue detail headings. All four layouts retain the same source-linked facts and unknowns. Which layout satisfies that organization without replacing the underlying evidence?

Select one answer, then check it to read the explanation. Enable JavaScript to select and check answers.

Practice question 3

Phase 1 of a synthetic repository exploration is complete. Its reviewed report identifies InvoiceView in src/invoice_view.py at revision r7, records a formatting dependency on MoneyFormatter in src/money.py at revision r4, and marks locale fallback behavior unresolved. Phase 2 will investigate that unresolved behavior in a fresh worker context. The orchestration contract requires phase-1 key findings and evidence in the worker’s initial context before dispatch; later file retrieval is unavailable in this exercise. Which preparation meets that contract?

Select one answer, then check it to read the explanation. Enable JavaScript to select and check answers.

Return to the human rubric and review your handoff

What this exercise establishes

These are original synthetic fixtures and an offline worksheet. The trace is a local representation of ordering, not an executed agent workflow. Version labels and digests identify the supplied bytes; mutable local files are not an authenticated or immutable evidence store.

No Claude, Agent SDK, MCP service, model or remote provider is executed. Passing the finite checks does not establish mastery, transfer effectiveness, exam readiness or correctness for other systems. These materials are not official exam questions or a full exam simulation.

Check the download bytes

SHA-256 identifies each supplied archive. It is a byte comparison, not a reasoning assessment or proof of authenticity.

Worksheet ZIP SHA-256
10a1ca327d7a182c3ce043b172aa9b4a1e5d09357ebbbf42758cb7bf1e2f962e
Worked answer ZIP SHA-256
5985850fddb9c78c7e1190cc491ef9c845a9358478484c9c7b015e35e6ad8a42

Continue the local practice