Update to Learning Commons

{
  "name": "k12-check-for-understanding",
  "description": "Creates a Check for Understanding (CFU) for a math topic or standard: 1–3 targeted formative items whose distractors come from documented student misconceptions, plus a teacher guide routing each likely response to a specific next instructional step. Load BEFORE asking any clarifying question about grade, standard, or topic. This skill is for math; for other subjects, help as you normally would and don't mention the skill to the user. Triggers on asks to create, write, generate, or design a check for understanding, formative check, quick check, or exit ticket for a math concept or standard, and on implicit signals like \"I need to see who got it before I move on.\" Not a quiz, test, or practice set — a CFU is read for what it reveals, not scored, so a numbered quiz or a bare answer key is a different job. Does not create a lesson (k12-lesson-plan-creation, which brings its own exit ticket), tier one (k12-lesson-differentiation), or prepare a teacher to teach one (k12-lesson-prep).\n",
  "included_files": [
    {
      "relative_path": "LICENSE",
      "size_in_bytes": 11358
    },
    {
      "relative_path": "NOTICE",
      "size_in_bytes": 320
    },
    {
      "relative_path": "references/learning-commons-kg.md",
      "size_in_bytes": 3410
    },
    {
      "relative_path": "references/math.md",
      "size_in_bytes": 6875
    },
    {
      "relative_path": "references/output.md",
      "size_in_bytes": 11809
    },
    {
      "relative_path": "references/verification.md",
      "size_in_bytes": 5170
    }
  ],
  "skill_md_contents": "---\nname: k12-check-for-understanding\ndescription: >\n  Creates a Check for Understanding (CFU) for a math topic or standard: 1–3 targeted formative items\n  whose distractors come from documented student misconceptions, plus a teacher guide routing each\n  likely response to a specific next instructional step. Load BEFORE asking any clarifying question\n  about grade, standard, or topic. This skill is for math; for other subjects, help as you normally\n  would and don't mention the skill to the user. Triggers on asks to create, write, generate, or\n  design a check for understanding, formative check, quick check, or exit ticket for a math concept\n  or standard, and on implicit signals like \"I need to see who got it before I move on.\" Not a\n  quiz, test, or practice set — a CFU is read for what it reveals, not scored, so a numbered quiz\n  or a bare answer key is a different job. Does not create a lesson (k12-lesson-plan-creation,\n  which brings its own exit ticket), tier one (k12-lesson-differentiation), or prepare a teacher to\n  teach one (k12-lesson-prep).\nlicense: \"Copyright 2026 Anthropic, PBC · Copyright 2026 Learning Commons · SPDX-License-Identifier: Apache-2.0. Complete terms in LICENSE and NOTICE.\"\n---\n\n# K-12 Check for Understanding\n\n1–3 formative items for a standard the teacher is about to teach or just taught: scoped to one\nlearning component within it, distractors drawn from documented student errors, plus a teacher guide\nsaying what each response reveals and what to do. A CFU is small on purpose — a five-minute read\nof where a class is, not an assessment event. \"The teacher\" is the user you're talking with.\n\n**Asset-based language everywhere**, in both files and every chat message. No \"weak,\" \"fragile,\"\n\"can't,\" \"behind,\" \"deficient,\" \"low,\" \"struggling,\" \"failing.\" Reframe with specificity, not\neuphemism: not \"fragile understanding of denominators\" but \"treats the denominator as a count of\nparts rather than the size of each part.\"\n\n---\n\n## Keeping the teacher posted\n\nKeep the teacher posted in a sentence or two as you go, in a teacher's words: the topic, the grade,\nwhat students will be asked to do, what they're getting. The machinery in this document — steps,\nlookups and connectors, learning components, the checks you run, file names and formats — stays\ninvisible; this applies to every message, progress notes included. \"I'll ground this in the\nstandard and the common errors for it, then build the check and a guide for reading responses\", not\n\"Reading the math build guidance required by Step 0.\"\n\n---\n\n## Step 0 — Route (silently)\n\n**Subject.** Math — arithmetic, fractions, ratio and proportion, geometry, algebra, functions,\nstatistics and probability, calculus, CCSS-M codes → **read `references/math.md` NOW**. Loading it is\nmandatory; designing items without it is a critical failure, and it is your skill instructions for\nthe turn. For ELA, science, or social studies, set this document aside and help as you normally would,\nwithout naming the skill. If the subject is ambiguous, ask in Step 1.\n\n**Connector.** Are Learning Commons KG tools (e.g. `find_standard_statement`) available? This decides\nStep 2; the skill works fully without them.\n\n---\n\n## Step 1 — Clarify\n\nAt most one question, and only when the answer changes the items:\n\n- **Grade**, when not stated or inferable — \"fractions\" spans several grades, and the standard\n  decides number values, representations, and what is on-grade. Ask before Step 2.\n- **Which standard**, when the topic maps to two plausible ones.\n\nEverything else defaults silently; never ask about item count, format, or rigor.\n\n---\n\n## Step 2 — Ground in the standard\n\n**Connected:** follow `references/learning-commons-kg.md` — call BEFORE designing anything; not\ncalling when connected is a critical failure.\n\n**Not connected:** work from the standard as you know it and add a footer to the teacher guide *\"The standard,\nprogression, and common errors here reflect general best practice.\"* with nothing\nabout what was or wasn't retrieved. Never invent KG citations or cite research you haven't seen.\n\n---\n\n## Step 3 — Select and confirm the learning-component focus\n\nA standard holds several separable ideas; scoping to one gives more actionable signal.\n\n1. Review **every** learning component returned in Step 2, if available.\n2. Select the **most assessable** — a genuine decision point in student understanding, probeable with\n   1–3 items, separating students at different levels of understanding.\n3. **State the selection and rationale, then stop and ask.** What the component asks students to do\n   in one plain sentence — a teacher's words, not the component's own wording — naming any constraint\n   it carries; one line on why; one direct question. **Do not proceed to Step 4 until the teacher\n   answers.**\n4. **If the teacher redirects** — a different component, or more than one — their choice becomes the\n   scope anchor. Restate it and its rationale in one line and scope every item to it.\n\nThen say in a sentence or two what you're about to do (*\"I'll pull the common errors research has\ndocumented for this standard, then build the check and a guide for reading the responses\"*) — teacher\nlanguage, no tool or file names.\n\n---\n\n## Step 4 — Build the check\n\nFollow the subject file's build section: coherence-map analysis, documented errors prioritized\nagainst the confirmed component, the misconception-to-stem map, item design by rigor, teacher-facing\nnotes.\n\n**Data-use guardrail.** Curriculum guidance and misconception data inform item design and next-step\nwording internally only. **No curriculum, publisher, or research-source attribution appears anywhere\nin either file or in any chat message.** Write original items; never reproduce retrieved\nstudent-facing text, contexts, or teacher notes verbatim.\n\n---\n\n## Step 5 — Output\n\nIn the turn Step 4 completes. Read `references/output.md` in full first, then write both files to\n`$OUTPUT_DIR`.\n\n---\n\n## Step 6 — Verify\n\n**Starts only once both files exist.** Do not open `references/verification.md` during Step 4 or 5:\nit is a gate on the written artifact, not build guidance, and reading it early turns it back into a\nchecklist you tick against what you meant to write.\n\nNow read it in full, read both files back from `$OUTPUT_DIR`, and run its two gates on every item —\nwhether more than one choice can be defended, and whether each distractor is a choice a real\nstudent's reasoning reaches. Fix what they find and rewrite the files. **Delivering a check that\nhasn't been through both gates is a critical failure.**\n\n---\n\n## Step 7 — Close\n\nAsk whether the check is what the teacher needs, offering at most two specific changes drawn from\nwhat this check is — a different rigor level, a second item on the same component, a\nconstructed-response version. Not \"let me know if you want changes.\" If they take one up, make it,\nrewrite both files, and ask again.\n\n**Plain language.** Never mention formats, file names, templates, how the documents were made, or the\nverification and what it changed: *\"Here's the check and a guide for reading the responses\"*, not\n*\"I've written two HTML files.\"*\n\n---\n\n## Cross-cutting checklist\n\n- [ ] Focus confirmed with the teacher first, and every item scoped to it\n- [ ] The check has 1–3 items, no more\n- [ ] Labeled rigor matches the standard, format follows the rigor table, and numbers and\n      representations respect the grade's range and the confirmed component's constraints\n- [ ] The coherence map grounds every prior-grade vs. on-grade route\n- [ ] One distractor is a prerequisite gap, one an on-grade confusion; no two rows route alike; every\n      next step names a sub-skill, representation, or task type\n- [ ] Every student-facing sentence sits inside its grade band (K–2 ~8 words, 3–5 ~12 with at most\n      one subordinate clause, 6–8 ~15), counted on the written stems and choices\n- [ ] Every figure carries `<title>`/`<desc>` with `role=\"img\"` and `aria-labelledby` in both files;\n      item text and choice order are identical between them\n- [ ] The student file has only prompts and choices/response space plus the name/date line; neither\n      file carries a provenance marker or names a source\n"
}