--- name: reconstruct-quantitative-solution-errors-test-transfer description: 'Trace a learner’s quantitative work to the earliest unsupported step, test competing error explanations, and design a fresh transfer task that measures reasoning rather than answer imitation. Use when a worked mathematics, statistics, accounting, economics, engineering, or other quantitative solution contains an unexplained error. An educator needs to distinguish a misconception from a slip before selecting an intervention. A learner has corrected an answer and needs a changed task to test transferable understanding.' compatibility: 'Instruction-only Agent Skill. Requires an Agent Skills-compatible client and access to the tools or systems named in the instructions. No executable scripts are bundled.' metadata: publisher: RichlyAI product: Amo.ng amo_id: AMO-S-000035 canonical_url: 'https://amo.ng/skills/reconstruct-quantitative-solution-errors-test-transfer' package_version: 1.0.0 source_type: prompt source_amo_id: AMO-P-000328 source_content_sha256: 42a0499c5836160ae076fcf87fc18dd9d927c8418e4693797d19dc12bed5cbfe source_version: 1.1.0 --- # Reconstruct Quantitative Solution Errors and Test Transfer ## When to use this Skill - A worked mathematics, statistics, accounting, economics, engineering, or other quantitative solution contains an unexplained error. - An educator needs to distinguish a misconception from a slip before selecting an intervention. - A learner has corrected an answer and needs a changed task to test transferable understanding. ## When not to use this Skill - Do not complete a restricted live assessment or provide a prohibited worked answer. - Do not diagnose learner intent, disability, effort, confidence, misconduct, or mastery from one artefact. - Do not assign an official grade or replace the authorized educator’s assessment process. ## Required inputs - Original quantitative problem, expected concepts, permitted methods, and assessment restrictions. - Learner work in its original order, including calculations, symbols, diagrams, explanations, and any assistance already received. - Authorized reference method, answer key, rubric, or source material. - Learning context, prior instruction, accommodation requirements, and the authorized educator or assessor. - Learner responses to diagnostic probes, repair attempts, and transfer task when those stages occur. Use [INPUT-TEMPLATE.md](assets/INPUT-TEMPLATE.md) to organize the supplied information. ## Required tools or access Published source tool designations: ChatGPT. These are source designations, not portable-client compatibility claims. Use only authorized access described in the inputs and source instructions. Do not claim access or inspection without evidence. ## Procedure Before starting, read [SOURCE.md](references/SOURCE.md), including the exact Prompt instructions and any ordered Workflow steps, review notes and carry-forward requirements. 1. Preserve the original order of the learner’s evidence and model the problem’s givens, target, constraints, and governing concepts. 2. Locate the earliest unsupported step rather than focusing only on the final wrong answer. 3. Form a small set of competing error explanations and identify evidence for, evidence against, and one discriminating question or micro-task for each. 4. Ask one probe at a time. Do not invent an absent learner response or proceed as if a probe succeeded. 5. Select the least-assistive repair consistent with the evidence and ask the learner to reconstruct the reasoning in their own notation or words. 6. Design a new transfer problem that preserves the target concept while changing numbers, representation, or context. Withhold its solution until the learner attempts it. 7. Record actual transfer reasoning, assistance used, error signatures, remaining uncertainty, and the next teaching action. Published expected output: A problem model, ordered step trace, earliest unsupported point, competing error-hypothesis table, discriminating probe, bounded corrective sequence, fresh transfer problem and rubric, observed transfer result when supplied, and educator handoff. ## Review gates Read and apply the source review instructions in [SOURCE.md](references/SOURCE.md) before acting. Keep proposed, unrun and executed actions distinct; preserve missing evidence and obtain the human approvals required by the source. ## Completion criteria Complete when the earliest unsupported step is evidence-linked; alternative explanations have been tested or remain explicitly uncertain; the learner has attempted a genuinely changed transfer problem; assistance is recorded; and the status reflects actual reasoning as Demonstrated, Demonstrated with help, Not demonstrated, Pending, or Blocked. Without learner responses, return a provisional diagnosis and test plan. Use the [completion checklist](references/COMPLETION-CHECKLIST.md) and record unresolved items. ## Limitations - Use only supplied work and responses; never fabricate omitted steps, checks, or learner evidence. - Preserve course rules, accessibility accommodations, privacy, and an equivalent way to demonstrate the target reasoning. - A correct answer alone is not proof of understanding; completion requires observable reasoning on a suitably changed problem. - The learner or authorized educator controls assessment, feedback, grading, and further intervention. This is an instruction-only package. Packaging does not execute work, grant access or independently verify the package. Existing public verification in amo.json describes only the identified published source resources. ## References - [Published source instructions](references/SOURCE.md) - [Completion checklist](references/COMPLETION-CHECKLIST.md) - [Input template](assets/INPUT-TEMPLATE.md) - [Rights](RIGHTS.md) - [Canonical Amo.ng Skill](https://amo.ng/skills/reconstruct-quantitative-solution-errors-test-transfer)