An agent organization structured as a company — 15+ departments, 125+ skills, each independently installable, citing the standards and regulators that settle the question. Runs in Claude Code and ChatGPT.
HTML agent-marketplacechatgpt-pluginchatgpt-skillsclaude-codeclaude-code-pluginclaude-pluginclaude-skillsmcporganization
Add the education vertical, and let a vertical bring a department
Claudedfd6fa1ea653
diff --git a/docs/DECISION-LOG.md b/docs/DECISION-LOG.mdindex c15a281..b886b42 100644--- a/docs/DECISION-LOG.md+++ b/docs/DECISION-LOG.md@@ -50,6 +50,7 @@ recorded rather than being deleted.| D36 | Where emitted vertical repositories live | ✅ Resolved || D37 | Authoritative sources as a catalog of references | ✅ Resolved || D38 | Running in ChatGPT as well as Claude Code | ✅ Resolved |+| D39 | A vertical that brings its own department | ✅ Resolved |---@@ -1350,3 +1351,42 @@ which is a different job from the skills, which describe how to do the work of a**The Codex category vocabulary is not published**, and Codex's own marketplace uses a small set ofbroad labels. Every department here is a business function, so all sixteen carry one conservativelabel rather than sixteen guesses. Revisit if a vocabulary is documented.++---++## D39. A vertical that brings its own department — ✅ Resolved++The education vertical was the second one built, and it immediately failed in a way industrial never+could. Industrial's skills belong to `operations` and `people`, which already exist; the generator+only ever had to add skills to departments it was already emitting. A curriculum function has no+home in a cross-industry core, because it is not a thinner version of anything every company has.++- **(a) Let a vertical declare a department, and have the generator do everything a new department+ requires.** ← **chosen**+- (b) File the education skills under an existing department — `product`, as curriculum-as-product,+ or `operations`. Needs no code, and misroutes every request that reaches it: a question about+ standards alignment loading a software product skill is the silent-collision failure+ `technology:skill-authoring` names.+- (c) Add `education` to the core, excluded from every other vertical. Puts a department nobody+ outside one industry wants into the default install, and inverts the exclusion list from a rare+ exception into the normal case.++**Resolution: (a).** A `[[department]]` entry in the vertical config, and the generator does what+the surface map requires of a new department because nobody is there to do it by hand: the plugin+manifest, the marketplace entry, the roster row, the surface block, and the charter — all in the+same emit, which is the rule the map states for a new department.++**Two adjacent things this exposed and fixed.**++Vertical skills were being copied as a lone `SKILL.md` rather than as a directory, so a skill's+`references/` never shipped. That went unnoticed while no vertical skill had any; the source catalog+gave education four that do, and the pointer would have shipped without its target.++The catalog could not reach a vertical's skills at all, because it resolved `department:skill` only+against `plugins/`. That was backwards: a vertical is where an outside authority matters *most*,+since the advice is industry-specific and therefore regulated by somebody. Both the checker and the+emitter now resolve against the core and every vertical.++**The emitted map drops the `verticals` and `sources` rows.** Their inputs stay upstream, so+downstream those rows would claim paths that are not there — a map that governs nothing, which is+the condition the authority column was added to eliminate.diff --git a/scripts/build-sources.py b/scripts/build-sources.pyindex c2bf3d2..7ba9873 100755--- a/scripts/build-sources.py+++ b/scripts/build-sources.py@@ -82,10 +82,19 @@ def targets():for _, entry in load_catalog():for ref in entry.get("skills", []):by_skill[ref].append(entry)+ import importlib.util+ spec = importlib.util.spec_from_file_location(+ "check_sources", os.path.join(os.path.dirname(os.path.abspath(__file__)), "check-sources.py"))+ checker = importlib.util.module_from_spec(spec)+ spec.loader.exec_module(checker)+out = {}for ref, entries in by_skill.items():department, skill = ref.split(":")- out[f"plugins/{department}/skills/{skill}/references/sources.md"] = render(ref, entries)+ path = checker.skill_path(department, skill)+ if not path:+ continue # check-sources.py reports this; emitting is not the place to fail on it+ out[os.path.join(os.path.dirname(path), "references", "sources.md")] = render(ref, entries)return out@@ -97,7 +106,8 @@ def main():wanted = targets()# A skill that loses its last source must lose the file too, or the catalog and the tree# disagree in the one direction a content comparison cannot see.- existing = {p for p in glob.glob("plugins/*/skills/*/references/sources.md")}+ existing = (set(glob.glob("plugins/*/skills/*/references/sources.md"))+ | set(glob.glob("verticals/*/skills/*/*/references/sources.md")))stale = sorted(existing - set(wanted))if args.check:diff --git a/scripts/build-vertical.py b/scripts/build-vertical.pyindex b629616..cc7996b 100755--- a/scripts/build-vertical.py+++ b/scripts/build-vertical.py@@ -34,6 +34,7 @@ import globimport jsonimport osimport re+import reimport shutilimport subprocessimport sys@@ -209,6 +210,44 @@ def collect(slug):return skills, context+def charter(entry, vertical_title):+ """The agent charter for a department a vertical brings with it."""+ name, title = entry["name"], entry["title"]+ return f"""---+name: {name}+description: {title}. Owns plugins/{name}/** and nothing else. Delegate work in this department's remit here.+---++# {title}++## Why this agent exists++The single owner of `plugins/{name}/**`. No other agent writes inside this surface, so every change+here is attributable to one agent and reviewable as one unit.++This department does not exist in the cross-industry core. It is brought by the {vertical_title}+vertical, because its remit is specific to that industry rather than a thinner version of something+every company has.++## Surface++Writes: `plugins/{name}/**`.+Reads: anything. Commits: nothing; the orchestrator is the sole committer.++## Standard++Skills in this department follow the conventions in `technology:skill-authoring`: the frontmatter+`name` equals the directory name, and the description carries both what the skill does and when to+reach for it.++## Verification this surface implies++- `python3 scripts/validate-skills.py` passes.+- `python3 scripts/check-provenance.py` passes — all content here is original.+- No change outside `plugins/{name}/**`. Needing one means coordinating with that surface's owner.+"""++def emit(slug, config, out):vertical, repo = config["vertical"], config["repo"]title = vertical["title"]@@ -219,6 +258,21 @@ def emit(slug, config, out):if unknown:die("exclude_departments names department(s) that are not on disk: " + ", ".join(sorted(unknown)))+ # A vertical may bring a department the core has no reason to carry. Industrial did not need+ # this — its skills belong to operations and people, which already exist. Education does: a+ # curriculum function is not a thinner version of anything in a cross-industry core, and+ # filing it under an existing department to avoid the feature would misroute every request.+ added = config.get("department", [])+ for entry in added:+ for field in ("name", "title", "description"):+ if not entry.get(field):+ die(f"verticals/{slug}: a [[department]] entry is missing {field}")+ if entry["name"] in departments:+ die(f"verticals/{slug}: department {entry['name']!r} already exists in the core — "+ f"add skills to it instead of declaring it")+ departments += [e["name"] for e in added]+ new_departments = {e["name"]: e for e in added}+skills, context = collect(slug)for (dept, skill) in list(skills) + list(context):if dept not in departments:@@ -242,9 +296,24 @@ def emit(slug, config, out):# Departments, with the vertical's material folded in.counts = {}for dept in departments:- src = os.path.join(ROOT, "plugins", dept)dst = os.path.join(out, "plugins", dept)- shutil.copytree(src, dst)+ if dept in new_departments:+ entry = new_departments[dept]+ os.makedirs(os.path.join(dst, "skills"))+ os.makedirs(os.path.join(dst, ".claude-plugin"))+ with open(os.path.join(dst, ".claude-plugin", "plugin.json"), "w",+ encoding="utf-8") as handle:+ json.dump({+ "name": dept,+ "description": entry["description"],+ "version": "1.0.0",+ "author": {"name": vertical.get("author", "Chris Brock")},+ "repository": repo["url"],+ "keywords": entry.get("keywords", []),+ }, handle, indent=2)+ handle.write("\n")+ else:+ shutil.copytree(os.path.join(ROOT, "plugins", dept), dst)for path in glob.glob(os.path.join(dst, "skills", "*", "SKILL.md")):skill = os.path.basename(os.path.dirname(path))fragment = context.get((dept, skill))@@ -255,9 +324,9 @@ def emit(slug, config, out):for (d, skill), path in skills.items():if d != dept:continue- target = os.path.join(dst, "skills", skill)- os.makedirs(target, exist_ok=True)- shutil.copy2(path, os.path.join(target, "SKILL.md"))+ # The whole directory, not just SKILL.md — a skill's `references/` is where its+ # sources live, and copying the file alone would ship the pointer without the target.+ shutil.copytree(os.path.dirname(path), os.path.join(dst, "skills", skill))manifest_path = os.path.join(dst, ".claude-plugin", "plugin.json")manifest = json.load(open(manifest_path, encoding="utf-8"))manifest["repository"] = repo["url"]@@ -271,6 +340,18 @@ def emit(slug, config, out):marketplace["name"] = namemarketplace["metadata"]["description"] = vertical.get("description", marketplace["metadata"]["description"])marketplace["plugins"] = [p for p in marketplace["plugins"] if p["name"] in departments]+ marketplace["plugins"] += [+ {+ "name": e["name"],+ "source": f"./plugins/{e['name']}",+ "description": e["description"],+ "version": "1.0.0",+ "author": {"name": vertical.get("author", "Chris Brock")},+ "keywords": e.get("keywords", []),+ "category": e.get("category", e["name"]),+ }+ for e in added+ ]os.makedirs(os.path.join(out, ".claude-plugin"))with open(os.path.join(out, ".claude-plugin", "marketplace.json"), "w", encoding="utf-8") as handle:json.dump(marketplace, handle, indent=2)@@ -278,8 +359,28 @@ def emit(slug, config, out):# The surface map, charters, and the checks that still apply downstream.os.makedirs(os.path.join(out, "docs"))- shutil.copy2(os.path.join(ROOT, "docs", "AGENT-SURFACES.md"), os.path.join(out, "docs"))+ surfaces = open(os.path.join(ROOT, "docs", "AGENT-SURFACES.md"), encoding="utf-8").read()shutil.copytree(os.path.join(ROOT, ".claude"), os.path.join(out, ".claude"))+ for entry in added:+ name = entry["name"]+ # A roster row, a surface block and a charter, in the same change — the rule the map states+ # for a new department, applied by the generator because nobody is there to apply it by hand.+ surfaces = surfaces.replace(+ "verticals builder installed proposes\n",+ f"{name:<20} builder installed autonomous\n", 1)+ surfaces = surfaces.replace(+ "```surface:repo-meta",+ f"```surface:{name}\nplugins/{name}/**\n```\n\n```surface:repo-meta", 1)+ with open(os.path.join(out, ".claude", "agents", f"{name}.md"), "w",+ encoding="utf-8") as handle:+ handle.write(charter(entry, title))+ # `verticals` and `sources` own nothing in an emitted repository — their inputs stay upstream —+ # so their rows would claim paths that are not here. Dropping them keeps the emitted map true.+ for gone in ("verticals", "sources"):+ surfaces = re.sub(rf"^{gone}\s+builder\s+installed\s+\w+\n", "", surfaces, flags=re.M)+ surfaces = re.sub(rf"```surface:{gone}\n.*?```\n\n", "", surfaces, flags=re.S)+ os.remove(os.path.join(out, ".claude", "agents", f"{gone}.md"))+ open(os.path.join(out, "docs", "AGENT-SURFACES.md"), "w", encoding="utf-8").write(surfaces)os.makedirs(os.path.join(out, "scripts"))for script in EMIT_SCRIPTS:shutil.copy2(os.path.join(ROOT, "scripts", script), os.path.join(out, "scripts"))diff --git a/scripts/check-sources.py b/scripts/check-sources.pyindex 6723b39..f6a378f 100755--- a/scripts/check-sources.py+++ b/scripts/check-sources.py@@ -29,6 +29,22 @@ ID = re.compile(r"^[a-z0-9]+(-[a-z0-9]+)*$")SKILL_REF = re.compile(r"^([a-z][a-z0-9-]*):([a-z][a-z0-9-]*)$")+def skill_path(department, skill):+ """Where a `department:skill` reference resolves, in the core or in any vertical.++ A vertical's skills are real skills — they are emitted into a shipped repository and an agent+ loads them the same way. Resolving only against `plugins/` would mean the catalog could never+ reach them, which is backwards: a vertical is where the outside authority matters most, because+ that is where the advice is industry-specific and therefore regulated by somebody.+ """+ core = f"plugins/{department}/skills/{skill}/SKILL.md"+ if os.path.exists(core):+ return core+ for path in sorted(glob.glob(f"verticals/*/skills/{department}/{skill}/SKILL.md")):+ return path+ return None++def load():"""Every source in the catalog, with the file it came from."""entries = []@@ -72,7 +88,7 @@ def validate(entries):if not match:problems.append(f"{where}: {ref!r} is not a department:skill reference")continue- if not os.path.exists(f"plugins/{match[1]}/skills/{match[2]}/SKILL.md"):+ if not skill_path(match[1], match[2]):problems.append(f"{where}: `{ref}` does not exist")return problems@@ -115,9 +131,12 @@ def pointers(entries):"""problems = []with_sources = {ref for _, e in entries for ref in e.get("skills", [])}- for path in sorted(glob.glob("plugins/*/skills/*/SKILL.md")):- parts = path.split(os.sep)- ref = f"{parts[1]}:{parts[3]}"+ every = ([(p, p.split(os.sep)[1], p.split(os.sep)[3])+ for p in glob.glob("plugins/*/skills/*/SKILL.md")]+ + [(p, p.split(os.sep)[3], p.split(os.sep)[4])+ for p in glob.glob("verticals/*/skills/*/*/SKILL.md")])+ for path, department, skill in sorted(every):+ ref = f"{department}:{skill}"mentions = "\n## Sources\n" in open(path, encoding="utf-8").read()if ref in with_sources and not mentions:problems.append(f"{path}: has catalog sources but no `## Sources` section — "diff --git a/sources/education.toml b/sources/education.tomlnew file mode 100644index 0000000..bbbbb7f--- /dev/null+++ b/sources/education.toml@@ -0,0 +1,160 @@+# Education standards, frameworks and primary-source collections.+#+# Almost nothing here is openly licensed, and that is the point of recording it. The public-domain+# core is the federal and national-collection material — Head Start, the Library of Congress, the+# National Archives. Common Core is usable only with its required attribution notice. Everything+# else is copyrighted and free to read.+#+# For anyone generating learning material the consequence is specific: align to a standard and cite+# its code, and write your own objective in your own words beside it. A paraphrase close enough to+# be useful is close enough to be a derivative.+#+# Reachability note. Several publishers here — thecorestandards.org, eclkc.ohs.acf.hhs.gov, loc.gov,+# naeyc.org — refuse automated requests from some networks with a 403. That is a bot filter rather+# than a dead link, and `check-sources.py --online` reports it as a warning for exactly that reason.+# The weekly workflow runs from a different network and is the real verification. Where a specific+# page could not be confirmed, the entry points at the publisher's canonical landing page rather+# than at a deep link nobody has fetched.++[[source]]+id = "ccss-math"+title = "Common Core State Standards — Mathematics"+publisher = "NGA Center for Best Practices & CCSSO"+url = "https://www.thecorestandards.org/Math/"+license = "attribution-required"+jurisdiction = "US"+authoritative_for = "The mathematics standards most US states adopted, by grade and domain, and the standard codes an alignment claim cites. The public license permits use with the required attribution notice — this is not public domain."+skills = ["education:standards-alignment", "education:assessment-design"]+checked = "2026-09-16"++[[source]]+id = "ccss-ela"+title = "Common Core State Standards — English Language Arts and Literacy"+publisher = "NGA Center for Best Practices & CCSSO"+url = "https://www.thecorestandards.org/ELA-Literacy/"+license = "attribution-required"+jurisdiction = "US"+authoritative_for = "Reading, writing, language and foundational-skills standards by grade — including the reading foundational skills that stop after grade 5, which is a common source of wrong alignment above it."+skills = ["education:standards-alignment", "education:learning-materials-design"]+checked = "2026-09-16"++[[source]]+id = "ccss-math-practice"+title = "Standards for Mathematical Practice"+publisher = "NGA Center for Best Practices & CCSSO"+url = "https://www.thecorestandards.org/Math/Practice/"+license = "attribution-required"+jurisdiction = "US"+authoritative_for = "The eight practice standards that run across every grade. Material claiming to build reasoning rather than procedure is aligning here, and usually does not say so."+skills = ["education:standards-alignment", "education:assessment-design"]+checked = "2026-09-16"++[[source]]+id = "ngss"+title = "Next Generation Science Standards"+publisher = "NGSS Lead States / Achieve"+url = "https://www.nextgenscience.org/search-standards"+license = "free-to-read"+jurisdiction = "US"+authoritative_for = "Science performance expectations by grade, each bundling a practice, a core idea and a crosscutting concept. Copyrighted, and the name is a registered trademark — cite the expectation code, never the wording."+skills = ["education:standards-alignment", "education:assessment-design"]+checked = "2026-09-16"++[[source]]+id = "ngss-appendices"+title = "NGSS appendices — practices, crosscutting concepts, core ideas"+publisher = "NGSS Lead States / Achieve"+url = "https://www.nextgenscience.org/resources/ngss-appendices"+license = "free-to-read"+jurisdiction = "US"+authoritative_for = "What three-dimensional learning means in practice, and why aligning to the disciplinary content alone misses two thirds of what a performance expectation asks."+skills = ["education:standards-alignment"]+checked = "2026-09-16"++[[source]]+id = "ncss-c3"+title = "C3 Framework for Social Studies State Standards"+publisher = "National Council for the Social Studies"+url = "https://www.socialstudies.org/standards/c3"+license = "free-to-read"+jurisdiction = "US"+authoritative_for = "The social studies inquiry arc — question, investigate, evaluate, act. It enumerates no facts, so an alignment claim to it is a claim about the shape of the task rather than its subject."+skills = ["education:standards-alignment"]+checked = "2026-09-16"++[[source]]+id = "head-start-elof"+title = "Head Start Early Learning Outcomes Framework, birth to five"+publisher = "US Department of Health and Human Services, Office of Head Start"+url = "https://eclkc.ohs.acf.hhs.gov/interactive-head-start-early-learning-outcomes-framework-ages-birth-five"+license = "public-domain-usgov"+jurisdiction = "US"+authoritative_for = "What is expected of children from birth to five across approaches to learning, social-emotional, language, cognition and physical development. The one comprehensive pre-K framework that is freely usable, where Common Core does not reach."+skills = ["education:early-childhood-practice", "education:standards-alignment"]+checked = "2026-09-16"++[[source]]+id = "naeyc-position-statements"+title = "NAEYC position statements, including Developmentally Appropriate Practice"+publisher = "National Association for the Education of Young Children"+url = "https://www.naeyc.org/resources/position-statements"+license = "free-to-read"+jurisdiction = "US"+authoritative_for = "What practice should look like for children under eight, and the statement most often cited when a task is challenged as too advanced. Not a standards document — it constrains method, not content."+skills = ["education:early-childhood-practice", "education:learning-materials-design"]+checked = "2026-09-16"++[[source]]+id = "nctm-standards"+title = "NCTM Principles and Standards and position statements"+publisher = "National Council of Teachers of Mathematics"+url = "https://www.nctm.org/standards-and-positions/principles-and-standards/"+license = "free-to-read"+jurisdiction = "US"+authoritative_for = "The process standards — problem solving, reasoning and proof, communication, connections, representation — and the profession's position on what mathematics instruction should do. Some material requires membership."+skills = ["education:assessment-design", "education:learning-materials-design"]+checked = "2026-09-16"++[[source]]+id = "nsta-positions"+title = "NSTA official positions"+publisher = "National Science Teaching Association"+url = "https://www.nsta.org/nstas-official-positions"+license = "free-to-read"+jurisdiction = "US"+authoritative_for = "The profession's positions on early childhood and elementary science instruction, including what inquiry-based means in practice at an age that cannot yet read the procedure."+skills = ["education:early-childhood-practice", "education:standards-alignment"]+checked = "2026-09-16"++[[source]]+id = "ila-resources"+title = "International Literacy Association position statements and briefs"+publisher = "International Literacy Association"+url = "https://www.literacyworldwide.org/get-resources"+license = "free-to-read"+jurisdiction = "global"+authoritative_for = "Research-informed positions on phonics, phonemic awareness, comprehension and writing instruction — the evidence base a literacy product's claims are judged against."+skills = ["education:learning-materials-design", "education:assessment-design"]+checked = "2026-09-16"++[[source]]+id = "loc-classroom"+title = "Library of Congress classroom materials"+publisher = "Library of Congress"+url = "https://www.loc.gov/classroom-materials/"+license = "public-domain-usgov"+jurisdiction = "US"+authoritative_for = "Primary sources usable in published material without a rights problem, with teaching context already attached. The first place to look before licensing an image."+skills = ["education:learning-materials-design"]+checked = "2026-09-16"++[[source]]+id = "nara-docsteach"+title = "DocsTeach"+publisher = "US National Archives and Records Administration"+url = "https://www.docsteach.org/"+license = "public-domain-usgov"+jurisdiction = "US"+authoritative_for = "Primary-source documents for US history material, freely usable, organized for classroom activities rather than for archivists."+skills = ["education:learning-materials-design"]+checked = "2026-09-16"diff --git a/verticals/README.md b/verticals/README.mdindex 7de87c6..ecb4e3b 100644--- a/verticals/README.md+++ b/verticals/README.md@@ -41,8 +41,28 @@ emit — the failure `build-readme.py` already guards against for departments.|---|---|| Add a skill | `skills/<department>/<skill>/SKILL.md`. Colliding with a core skill's name fails the emit. || Extend a core skill | `context/<department>/<skill>.md`, spliced in ahead of the skill's `## Never` block. |+| Add a department | A `[[department]]` entry in the config. || Drop a department | `[emit] exclude_departments` in the config. |+**A vertical may bring a department the core has no reason to carry.** Industrial does not: its+skills belong to `operations` and `people`, which already exist. Education does — a curriculum+function is not a thinner version of anything in a cross-industry core, and filing it under an+existing department to avoid the feature would misroute every request that reached it.++```toml+[[department]]+name = "education"+title = "Education"+description = "Standards alignment, learning-materials design, assessment construction."+keywords = ["curriculum", "standards"]+```++The generator does what the surface map requires of a new department, because nobody is there to do+it by hand: it writes the plugin manifest, adds the marketplace entry, inserts the roster row and+surface block into the emitted map, and generates the charter. It also drops the `verticals` and+`sources` rows from the emitted map — their inputs stay upstream, so downstream those rows would+claim paths that are not there.+It may **not** edit a core skill in place. A core skill that is wrong for every industry is wrong inthe core; fixing it there means every vertical gets the fix on its next emit, which is the wholereason for generating rather than forking.diff --git a/verticals/education/skills/education/assessment-design/SKILL.md b/verticals/education/skills/education/assessment-design/SKILL.mdnew file mode 100644index 0000000..33865aa--- /dev/null+++ b/verticals/education/skills/education/assessment-design/SKILL.md@@ -0,0 +1,96 @@+---+name: assessment-design+description: Builds questions and assessments that measure what they claim — matching item type to the thing being assessed, writing distractors that diagnose, and telling a question that measures learning from one that measures reading speed or test-taking. Use this to write or review assessment items, design a unit check or diagnostic, work out why scores do not match classroom performance, or decide how much evidence a claim about mastery actually needs.+---++# Assessment design++Every question measures something. The work is making sure it measures the thing you meant, because+the alternatives — reading speed, familiarity with the format, willingness to guess — are always+available and are usually easier for the student to use.++## Name the inference before writing the item++An assessment is an argument: the student did this, therefore they can do that. The argument is+where assessments fail, and it fails silently.++Write down the claim first — "can decompose a two-digit number into tens and ones" — then ask what+performance would be evidence for it, and what performance would be evidence *against*. An item+that a student who lacks the skill can still get right is not evidence. An item that a student who+has the skill can still get wrong, for reasons unrelated to it, is worse: it produces a false+negative that gets acted on.++The most common unrelated reason is reading. Any item whose stem is harder to read than the skill is+to perform has quietly become a reading assessment.++## Match the item type to the claim++- **Selected response** (multiple choice, matching, true/false) is efficient and can only ever+ provide evidence of recognition. A student who can recognize the correct answer cannot be assumed+ to produce it.+- **Constructed response** shows the path, which is what makes partial understanding visible. It+ costs scoring time and needs a rubric written before the responses arrive, not after.+- **Performance tasks** are the only honest evidence for anything described as applying,+ investigating or designing — most NGSS performance expectations and every C3 inquiry, for+ instance, cannot be assessed by selected response at all.++Mismatch is the usual failure: a standard describing explanation assessed by a multiple-choice item,+which measures whether the student can pick an explanation someone else wrote.++## Distractors are the diagnostic++In a well-built multiple-choice item, each wrong answer is the result of a specific, predictable+error. Then the pattern of wrong answers says what to reteach, and the item earns its place.++Distractors that are merely wrong — a random number, an obviously absurd option — turn a four-option+item into a two-option one and tell you nothing beyond right or wrong.++The mechanical tells of a weak item, all of which students learn to exploit long before they learn+the content:++- The longest or most qualified option is correct.+- One option is grammatically inconsistent with the stem.+- "All of the above" appears, and is usually correct.+- Two options are synonyms, so neither can be right.+- The correct answer repeats wording from the stem.++## One item is not evidence++A single item carries noise — a misread word, a slip, a lucky guess — that swamps the signal for any+individual student. Inferring mastery from one response is the most common measurement error in+classroom material, and the most consequential, because it gets recorded.++Several items per claim, varied in surface form so that recognition of the format is not what is+being measured. If a claim is worth recording against a student's name, it is worth three items.++Spacing matters as much as quantity: performance on a skill immediately after it is taught measures+something closer to short-term recall than to learning. The same item a fortnight later measures+more.++## Formative and summative are different products++Formative assessment exists to change what happens next, which means it must be quick, frequent,+low-stakes, and read immediately. An assessment that takes a week to score cannot be formative+whatever it is called.++Summative assessment exists to record a judgment, which means it must be defensible: enough items,+a rubric written in advance, and conditions that were the same for everyone.++Material sold as practice is usually formative in function and summative in appearance, which is+worth stating plainly to a buyer. `education:learning-materials-design` covers the practice+sequence this sits inside.++## Sources++`references/sources.md` in this skill lists the outside authorities that settle the questions+here — what each one is authoritative for, and what you may do with it. Check them before+answering on anything they cover, and cite what you used. Most are free to read and not free+to reproduce; the use note on each is binding.++## Never++- Write an item before naming the claim it is evidence for.+- Assess an explanation standard with a recognition item.+- Build distractors that are wrong without being diagnostic.+- Record mastery from a single response.+- Make the stem harder to read than the skill is to perform.diff --git a/verticals/education/skills/education/assessment-design/references/sources.md b/verticals/education/skills/education/assessment-design/references/sources.mdnew file mode 100644index 0000000..54b7e15--- /dev/null+++ b/verticals/education/skills/education/assessment-design/references/sources.md@@ -0,0 +1,49 @@+# Sources — `education:assessment-design`++<!-- Generated by scripts/build-sources.py from sources/*.toml. Do not edit. -->++Check these before answering on anything they cover, and cite what you used. The use note on each one is binding: most of what a professional cites is free to read and not free to reproduce.++## Common Core State Standards — Mathematics++NGA Center for Best Practices & CCSSO · US · free to use with attribution — credit the publisher++<https://www.thecorestandards.org/Math/>++**Authoritative for:** The mathematics standards most US states adopted, by grade and domain, and the standard codes an alignment claim cites. The public license permits use with the required attribution notice — this is not public domain.++## International Literacy Association position statements and briefs++International Literacy Association · global · **read and cite only — copyrighted, do not reproduce**++<https://www.literacyworldwide.org/get-resources>++**Authoritative for:** Research-informed positions on phonics, phonemic awareness, comprehension and writing instruction — the evidence base a literacy product's claims are judged against.++## NCTM Principles and Standards and position statements++National Council of Teachers of Mathematics · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nctm.org/standards-and-positions/principles-and-standards/>++**Authoritative for:** The process standards — problem solving, reasoning and proof, communication, connections, representation — and the profession's position on what mathematics instruction should do. Some material requires membership.++## Next Generation Science Standards++NGSS Lead States / Achieve · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nextgenscience.org/search-standards>++**Authoritative for:** Science performance expectations by grade, each bundling a practice, a core idea and a crosscutting concept. Copyrighted, and the name is a registered trademark — cite the expectation code, never the wording.++## Standards for Mathematical Practice++NGA Center for Best Practices & CCSSO · US · free to use with attribution — credit the publisher++<https://www.thecorestandards.org/Math/Practice/>++**Authoritative for:** The eight practice standards that run across every grade. Material claiming to build reasoning rather than procedure is aligning here, and usually does not say so.++---++Sources are maintained in `sources/` upstream, not here. If one is wrong, out of date, or missing, fix it there — this file is regenerated and an edit to it is lost.diff --git a/verticals/education/skills/education/early-childhood-practice/SKILL.md b/verticals/education/skills/education/early-childhood-practice/SKILL.mdnew file mode 100644index 0000000..a0f928d--- /dev/null+++ b/verticals/education/skills/education/early-childhood-practice/SKILL.md@@ -0,0 +1,89 @@+---+name: early-childhood-practice+description: Judges whether material for children under six is developmentally plausible — what is typical at each age, why pre-K is not a thinner kindergarten, and which frameworks apply where Common Core does not. Use this to design or review pre-K material, decide whether a task is reasonable for an age, choose the right framework for a pre-K alignment claim, or push back on material that pushes school-age expectations downward.+---++# Early childhood practice++Pre-K is not grade zero. Material for three- to five-year-olds fails most often by being a smaller+version of kindergarten work — the same task, fewer items, larger print — which is a different+mistake from being too hard.++## There is no Common Core below kindergarten++This is the first thing to establish, because the alignment claim printed on most pre-K material is+either wrong or vague.++What does apply:++- **State early learning standards**, which exist in every state and vary substantially in structure+ and in what they expect at each age.+- **The Head Start Early Learning Outcomes Framework**, which covers birth to five across+ approaches to learning, social-emotional, language and literacy, cognition, and perceptual and+ physical development. It is a federal work and freely usable.+- **NAEYC's position on developmentally appropriate practice**, which is not a standards document at+ all — it is a statement of what practice should look like, and is the thing most often cited when+ a task is challenged.++A claim of "Common Core aligned" on pre-K material is not defensible.+`education:standards-alignment` covers alignment above kindergarten.++## Developmentally appropriate is a range, not a line++The useful question is not whether a four-year-old can do a task. Some can. It is whether a typical+four-year-old can do it without the experience becoming one of failing.++Ranges worth holding, with the caveat that variation at these ages is wide and normal:++- **Fine motor.** Tripod grasp is still consolidating at four. Tracing within narrow lines,+ small-square handwriting and precision cutting are ahead of most of them, and repeated failure at+ a motor task teaches avoidance of the subject it was attached to.+- **Attention.** Sustained focus on an adult-directed task runs in single-digit minutes. A page that+ takes twenty minutes is not one task, it is four, and it should be designed as four.+- **Counting and cardinality.** Reciting number names comes well before understanding that the last+ number said is how many there are. Material that assumes the second because the first is fluent+ misreads the child.+- **Letters.** Recognition, naming, sound and formation are four separate skills that develop at+ different rates. Name-writing typically precedes general letter formation, because the motivation+ is different.+- **Symbolic reasoning.** A picture standing for a thing is not automatic at three and is usually+ secure by five. Worksheets depend on it entirely.++## Play is the method, which constrains the format++The consensus position across the early childhood field is that young children learn through play,+exploration and interaction — not that play is a reward after the work.++This is uncomfortable for anyone producing worksheets, and the honest conclusion is that **the+worksheet is a supporting format at this age, not the primary one**. Material that acknowledges this+is better material: a page that prompts an activity, a conversation or a hunt around the room does+more than the same page asking for tracing, and it is what a knowledgeable reviewer looks for.++Where a page is the deliverable, the ones that work at this age tend to be open rather than+closed — draw, sort, match, find — with no single correct answer to get wrong.++## Pushdown is the failure mode to name++Academic expectations have migrated downward, so kindergarten now carries work that was first grade+a generation ago, and pre-K material follows. The pressure is commercial: parents buy what looks+rigorous, and "gets them ready" sells.++Being explicit about this is part of doing the work honestly. A pre-K page asking for sentence+writing is not ambitious, it is misaimed, and the child who cannot do it concludes something about+themselves rather than about the page. When a buyer asks for material a year ahead, that is a real+request and it deserves a straight answer about what it will cost the child.++## Sources++`references/sources.md` in this skill lists the outside authorities that settle the questions+here — what each one is authoritative for, and what you may do with it. Check them before+answering on anything they cover, and cite what you used. Most are free to read and not free+to reproduce; the use note on each is binding.++## Never++- Claim Common Core alignment for material below kindergarten.+- Treat a pre-K task as a kindergarten task with fewer items.+- Require fine-motor precision the age has not typically reached.+- Design a single page that needs more sustained attention than the age sustains.+- Present a worksheet as the primary way a three-year-old should be learning.diff --git a/verticals/education/skills/education/early-childhood-practice/references/sources.md b/verticals/education/skills/education/early-childhood-practice/references/sources.mdnew file mode 100644index 0000000..f202e69--- /dev/null+++ b/verticals/education/skills/education/early-childhood-practice/references/sources.md@@ -0,0 +1,33 @@+# Sources — `education:early-childhood-practice`++<!-- Generated by scripts/build-sources.py from sources/*.toml. Do not edit. -->++Check these before answering on anything they cover, and cite what you used. The use note on each one is binding: most of what a professional cites is free to read and not free to reproduce.++## Head Start Early Learning Outcomes Framework, birth to five++US Department of Health and Human Services, Office of Head Start · US · public domain (US government) — quote freely++<https://eclkc.ohs.acf.hhs.gov/interactive-head-start-early-learning-outcomes-framework-ages-birth-five>++**Authoritative for:** What is expected of children from birth to five across approaches to learning, social-emotional, language, cognition and physical development. The one comprehensive pre-K framework that is freely usable, where Common Core does not reach.++## NAEYC position statements, including Developmentally Appropriate Practice++National Association for the Education of Young Children · US · **read and cite only — copyrighted, do not reproduce**++<https://www.naeyc.org/resources/position-statements>++**Authoritative for:** What practice should look like for children under eight, and the statement most often cited when a task is challenged as too advanced. Not a standards document — it constrains method, not content.++## NSTA official positions++National Science Teaching Association · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nsta.org/nstas-official-positions>++**Authoritative for:** The profession's positions on early childhood and elementary science instruction, including what inquiry-based means in practice at an age that cannot yet read the procedure.++---++Sources are maintained in `sources/` upstream, not here. If one is wrong, out of date, or missing, fix it there — this file is regenerated and an edit to it is lost.diff --git a/verticals/education/skills/education/learning-materials-design/SKILL.md b/verticals/education/skills/education/learning-materials-design/SKILL.mdnew file mode 100644index 0000000..8cd8002--- /dev/null+++ b/verticals/education/skills/education/learning-materials-design/SKILL.md@@ -0,0 +1,115 @@+---+name: learning-materials-design+description: Designs worksheets, workbooks and practice material that actually teach — sequencing difficulty, managing cognitive load, writing instructions a child can follow unaided, and building answer keys that hold up. Use this to design or fix a worksheet or workbook, decide how much practice a skill needs, structure a page so it can be completed without an adult reading it aloud, or diagnose material that looks right and does not work.+---++# Learning materials design++Most bad worksheets are not wrong. They are unusable: the instruction assumes a reading level above+the skill being practiced, the page asks the student to hold three things in mind at once, or the+difficulty jumps in a place the author could not see because the author already knows the answer.++## The instruction is the hardest part of the page++A worksheet is used by a child, often without an adult reading it to them, and frequently by a+parent who did not see the lesson. The instruction has to survive that.++Two constraints that conflict and have to be resolved deliberately:++- **The instruction's reading level must sit below the skill being practiced.** A second-grade math+ page whose directions need third-grade reading is a reading test with numbers on it. The child who+ fails it may have no trouble with the math.+- **The instruction must be complete.** "Solve" does not say whether to show work, what form the+ answer takes, or what to do with the remainder.++Where the two cannot both be met, the answer is usually a worked example rather than more words. One+solved item at the top carries more instruction than a paragraph, and it is read.++## Cognitive load is the budget++Working memory is the scarce resource, and a page spends it on three things at once: the skill being+learned, the format it is presented in, and everything incidental — decoration, unfamiliar layout, a+font that makes a `9` look like a `g`.++Only the first is worth spending on. The practical consequences:++- **Change one thing at a time.** A page that introduces a new skill *and* a new answer format+ measures neither.+- **Keep the format stable across a section.** Familiarity with the layout is load the student stops+ paying after the first page, and resetting it is a cost with no return.+- **Decoration competes with content.** Illustration that carries meaning earns its space;+ illustration that fills it costs attention, most sharply for the students who have least to spare.++## Difficulty rises in steps somebody checked++The author knows the answer, which makes difficulty jumps invisible from the inside. The reliable+method is mechanical: list the sub-skills the final task requires, order them by dependency, and+confirm every item exercises only sub-skills already introduced.++Where practice sits in a sequence changes what it should look like:++1. **Acquisition** — heavily scaffolded, worked examples adjacent, immediate feedback possible.+ Errors here are information, so the page should make them visible rather than punish them.+2. **Fluency** — scaffolding removed, volume up, the same skill in the same form. This is the stage+ most commercial material skips, and the reason a student who "understands it" cannot do it in+ March.+3. **Generalization** — the skill in unfamiliar wrappers, mixed with earlier skills, in contexts the+ examples did not cover. This is where transfer is tested and where spaced review belongs.++Mixing stages on one page is common and usually a mistake: a generalization item among acquisition+items reads as unfair, and an acquisition item among fluency items reads as filler.++## The answer key is part of the product++An answer key that is wrong is worse than no key, because it is trusted. It is also where multiple+valid answers get discovered — the page that admits two readings, the drawing task with no+determinate result, the word problem whose units were never stated.++Build the key by solving the page cold, from the printed instruction, without reference to what was+intended. Every item that is ambiguous when solved that way is ambiguous to the student, and the fix+belongs on the page rather than in the key.++For anything open-ended, the key states what makes an answer acceptable rather than pretending there+is one answer. A parent needs that more than the child does.++## Print is a constraint, not a format++Material that will be printed or sold as a file has limits that a screen mockup hides: margin loss+on a home printer, greyscale rendering of a color-coded task, and the fact that nobody can zoom.++Design in black and white first and add color as reinforcement rather than as the carrier of+meaning. A task that requires distinguishing red from green fails in greyscale and fails for the+significant minority of readers who cannot make that distinction anyway.++`marketing:visual-content` covers the design craft; what is specific here is that the constraint is+a child's hand, a home printer and no adult present.++## Tooling++Layout for print material is a document problem before it is a design one. InDesign, Affinity+Publisher, Canva and similar all produce a usable page; what determines whether a hundred-page+workbook stays maintainable is whether the layout is driven by styles and templates rather than+placed by hand, because the correction always arrives after page sixty.++Generated material — the same task shape across many values — is worth scripting rather than+copying. A script that emits items from a parameter set makes the difficulty sequence explicit and+regenerable, and it is the only way a mistake in item construction gets fixed once rather than+eighty times.++For anything sold, keep the source and the output separate and treat the PDF as a build artifact.+The alternative is a product nobody can correct.++## Sources++`references/sources.md` in this skill lists the outside authorities that settle the questions+here — what each one is authoritative for, and what you may do with it. Check them before+answering on anything they cover, and cite what you used. Most are free to read and not free+to reproduce; the use note on each is binding.++## Never++- Write an instruction that reads above the level of the skill being practiced.+- Introduce a new skill and a new answer format on the same page.+- Build the answer key from what you intended rather than from what the page says.+- Carry meaning in color alone.+- Ship a difficulty sequence nobody walked through from the student's side.diff --git a/verticals/education/skills/education/learning-materials-design/references/sources.md b/verticals/education/skills/education/learning-materials-design/references/sources.mdnew file mode 100644index 0000000..dd0cd78--- /dev/null+++ b/verticals/education/skills/education/learning-materials-design/references/sources.md@@ -0,0 +1,57 @@+# Sources — `education:learning-materials-design`++<!-- Generated by scripts/build-sources.py from sources/*.toml. Do not edit. -->++Check these before answering on anything they cover, and cite what you used. The use note on each one is binding: most of what a professional cites is free to read and not free to reproduce.++## Common Core State Standards — English Language Arts and Literacy++NGA Center for Best Practices & CCSSO · US · free to use with attribution — credit the publisher++<https://www.thecorestandards.org/ELA-Literacy/>++**Authoritative for:** Reading, writing, language and foundational-skills standards by grade — including the reading foundational skills that stop after grade 5, which is a common source of wrong alignment above it.++## DocsTeach++US National Archives and Records Administration · US · public domain (US government) — quote freely++<https://www.docsteach.org/>++**Authoritative for:** Primary-source documents for US history material, freely usable, organized for classroom activities rather than for archivists.++## International Literacy Association position statements and briefs++International Literacy Association · global · **read and cite only — copyrighted, do not reproduce**++<https://www.literacyworldwide.org/get-resources>++**Authoritative for:** Research-informed positions on phonics, phonemic awareness, comprehension and writing instruction — the evidence base a literacy product's claims are judged against.++## Library of Congress classroom materials++Library of Congress · US · public domain (US government) — quote freely++<https://www.loc.gov/classroom-materials/>++**Authoritative for:** Primary sources usable in published material without a rights problem, with teaching context already attached. The first place to look before licensing an image.++## NAEYC position statements, including Developmentally Appropriate Practice++National Association for the Education of Young Children · US · **read and cite only — copyrighted, do not reproduce**++<https://www.naeyc.org/resources/position-statements>++**Authoritative for:** What practice should look like for children under eight, and the statement most often cited when a task is challenged as too advanced. Not a standards document — it constrains method, not content.++## NCTM Principles and Standards and position statements++National Council of Teachers of Mathematics · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nctm.org/standards-and-positions/principles-and-standards/>++**Authoritative for:** The process standards — problem solving, reasoning and proof, communication, connections, representation — and the profession's position on what mathematics instruction should do. Some material requires membership.++---++Sources are maintained in `sources/` upstream, not here. If one is wrong, out of date, or missing, fix it there — this file is regenerated and an edit to it is lost.diff --git a/verticals/education/skills/education/standards-alignment/SKILL.md b/verticals/education/skills/education/standards-alignment/SKILL.mdnew file mode 100644index 0000000..f0e3737--- /dev/null+++ b/verticals/education/skills/education/standards-alignment/SKILL.md@@ -0,0 +1,106 @@+---+name: standards-alignment+description: Aligns learning material to the standard it claims — choosing the right framework for the grade and subject, citing standards by code rather than reproducing their text, and telling a real alignment from a decorative one. Use this to align a worksheet, lesson or workbook to Common Core, NGSS or a state framework, to check an alignment claim before publishing, or to decide what may lawfully be printed on the cover.+---++# Standards alignment++An alignment claim is a promise to a buyer who cannot verify it quickly. Most published alignment is+decorative — a code printed in a corner, chosen after the worksheet was written, that the task does+not actually assess. It survives because nobody checks, until a district curriculum reviewer does.++## Align to the verb, not the topic++A standard is written as a performance: what the student does, to what, under what conditions. The+topic is the least specific part of it and the part most alignment stops at.++`3.NF.A.1` is about fractions. It is specifically about *understanding a fraction as a quantity+formed by parts of a whole* — so a page of shaded-circle identification aligns, and a page of+fraction addition drills does not, however much both are "fractions in third grade."++The test to apply before printing a code: **could a student complete this task correctly without+doing what the standard describes?** If yes, the alignment is decorative. A word problem solvable by+guessing from the answer format assesses test-taking, and it is aligned to nothing.++## Know which framework applies++Grade and subject do not determine the framework on their own, and the mismatch is a common source+of wrong claims:++- **Common Core** covers mathematics and English language arts, K-12. It is adopted by most but not+ all states, and several adopting states have renamed and amended it — a "Common Core aligned"+ claim in a state that rewrote it is at best imprecise.+- **NGSS** covers science, and is organized as performance expectations that bundle a practice, a+ core idea and a crosscutting concept. Aligning to the disciplinary content alone misses two+ thirds of what the expectation asks for.+- **C3** covers social studies, and is an inquiry arc rather than a content list. It does not+ enumerate facts, so "aligned to C3" is a claim about the shape of the task, not its subject.+- **Pre-K has no Common Core.** Nothing below kindergarten is covered, and an alignment claim there+ is either to a state early learning standard, to Head Start's framework, or to nothing. Treat+ pre-K as its own problem — `education:early-childhood-practice` holds it.++State standards exist alongside all of these, and for a product sold into one state they are the+binding set. `education:learning-materials-design` covers what changes when the material is for a+named district rather than a general market.++## Cite the code, never the text++This is where alignment becomes a legal question rather than an editorial one, and the distinction+is sharper than most publishers treat it.++**Citing a standard by its code is not reproduction.** `4-PS3-2`, `RL.5.2`, `2.MD.C.7` are+identifiers. Using them to say what a page addresses is normal, expected, and safe.++**Reproducing the standard's wording is a different act**, and whether it is permitted depends on+the publisher:++- Common Core carries a public license that permits use **with its required attribution notice**.+ It is not public domain, and dropping the notice is the common failure.+- NGSS is copyrighted, and "Next Generation Science Standards" is a registered trademark — which+ constrains the alignment claim itself, not only the text.+- State frameworks vary. Many are state government works and freely usable; some are licensed from+ a vendor and are not.++Paraphrasing to avoid the question does not avoid it, and a paraphrase that stays close enough to be+useful is close enough to be a derivative. Write your own learning objective in your own words, and+cite the code beside it. That is both safer and clearer to a teacher.++`references/sources.md` carries the publisher and license class for each framework.++## An alignment document is the deliverable++For anything sold to a school, the alignment is an artifact, not an annotation. It lists, per page+or per lesson: the code, the text of *your* objective, the task that provides evidence, and where a+reviewer would look to verify it.++Building it after the material is written is what produces decorative alignment. Building it first+makes it a specification: each row is a task somebody then has to write.++## Tooling++Standards data is available as structured identifiers more often than publishers realize. Common+Core and NGSS codes are stable and enumerable, and a spreadsheet keyed on them beats prose in a+document — it makes coverage gaps visible, which is the question a reviewer asks first.++Curriculum mapping tools — Chalk, Atlas, Eduplanet21, Kiddom and similar — hold the alignment+document and the sequence together, and earn their place once the same material is sold into+several states with different frameworks. Below that, a maintained spreadsheet with one row per+standard and one column per unit is the same artifact without the subscription.++Whatever holds it, keep the standard's code and your objective in separate fields. Merging them is+how the standard's wording ends up in the product.++## Sources++`references/sources.md` in this skill lists the outside authorities that settle the questions+here — what each one is authoritative for, and what you may do with it. Check them before+answering on anything they cover, and cite what you used. Most are free to read and not free+to reproduce; the use note on each is binding.++## Never++- Print a standard code chosen after the task was written.+- Claim alignment to a framework the state in question does not use.+- Reproduce a standard's wording without the license notice its publisher requires.+- Align to the topic in the standard and ignore the performance it describes.+- Claim Common Core alignment for anything below kindergarten.diff --git a/verticals/education/skills/education/standards-alignment/references/sources.md b/verticals/education/skills/education/standards-alignment/references/sources.mdnew file mode 100644index 0000000..283ed4f--- /dev/null+++ b/verticals/education/skills/education/standards-alignment/references/sources.md@@ -0,0 +1,73 @@+# Sources — `education:standards-alignment`++<!-- Generated by scripts/build-sources.py from sources/*.toml. Do not edit. -->++Check these before answering on anything they cover, and cite what you used. The use note on each one is binding: most of what a professional cites is free to read and not free to reproduce.++## C3 Framework for Social Studies State Standards++National Council for the Social Studies · US · **read and cite only — copyrighted, do not reproduce**++<https://www.socialstudies.org/standards/c3>++**Authoritative for:** The social studies inquiry arc — question, investigate, evaluate, act. It enumerates no facts, so an alignment claim to it is a claim about the shape of the task rather than its subject.++## Common Core State Standards — English Language Arts and Literacy++NGA Center for Best Practices & CCSSO · US · free to use with attribution — credit the publisher++<https://www.thecorestandards.org/ELA-Literacy/>++**Authoritative for:** Reading, writing, language and foundational-skills standards by grade — including the reading foundational skills that stop after grade 5, which is a common source of wrong alignment above it.++## Common Core State Standards — Mathematics++NGA Center for Best Practices & CCSSO · US · free to use with attribution — credit the publisher++<https://www.thecorestandards.org/Math/>++**Authoritative for:** The mathematics standards most US states adopted, by grade and domain, and the standard codes an alignment claim cites. The public license permits use with the required attribution notice — this is not public domain.++## Head Start Early Learning Outcomes Framework, birth to five++US Department of Health and Human Services, Office of Head Start · US · public domain (US government) — quote freely++<https://eclkc.ohs.acf.hhs.gov/interactive-head-start-early-learning-outcomes-framework-ages-birth-five>++**Authoritative for:** What is expected of children from birth to five across approaches to learning, social-emotional, language, cognition and physical development. The one comprehensive pre-K framework that is freely usable, where Common Core does not reach.++## NGSS appendices — practices, crosscutting concepts, core ideas++NGSS Lead States / Achieve · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nextgenscience.org/resources/ngss-appendices>++**Authoritative for:** What three-dimensional learning means in practice, and why aligning to the disciplinary content alone misses two thirds of what a performance expectation asks.++## NSTA official positions++National Science Teaching Association · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nsta.org/nstas-official-positions>++**Authoritative for:** The profession's positions on early childhood and elementary science instruction, including what inquiry-based means in practice at an age that cannot yet read the procedure.++## Next Generation Science Standards++NGSS Lead States / Achieve · US · **read and cite only — copyrighted, do not reproduce**++<https://www.nextgenscience.org/search-standards>++**Authoritative for:** Science performance expectations by grade, each bundling a practice, a core idea and a crosscutting concept. Copyrighted, and the name is a registered trademark — cite the expectation code, never the wording.++## Standards for Mathematical Practice++NGA Center for Best Practices & CCSSO · US · free to use with attribution — credit the publisher++<https://www.thecorestandards.org/Math/Practice/>++**Authoritative for:** The eight practice standards that run across every grade. Material claiming to build reasoning rather than procedure is aligning here, and usually does not say so.++---++Sources are maintained in `sources/` upstream, not here. If one is wrong, out of date, or missing, fix it there — this file is regenerated and an edit to it is lost.diff --git a/verticals/education/vertical.toml b/verticals/education/vertical.tomlnew file mode 100644index 0000000..60bf101--- /dev/null+++ b/verticals/education/vertical.toml@@ -0,0 +1,42 @@+# The education vertical: curriculum-aligned learning materials.+#+# The second vertical, and the one that made the generator support a department the core does not+# have. Industrial's skills belong to operations and people, which already exist. A curriculum+# function is not a thinner version of anything in a cross-industry core, and filing it under an+# existing department to avoid building the feature would misroute every request that reached it.++[vertical]+slug = "education"+name = "headcount-education"+title = "Education"+tagline = "An agent organization for people who make learning materials — the full cross-industry core, plus standards alignment, materials design and assessment."+description = "An agent organization for curriculum publishers, learning-materials authors and education programs. Each department installs independently so only the relevant function loads."+keywords = ["education", "curriculum", "learning-materials", "standards-alignment"]+rationale = """+Making learning materials is a publishing business with a compliance problem attached. The core+covers the publishing side — positioning, pricing, content operations, unit economics — and says+nothing about the part that decides whether the material is any good or lawful to sell: whether it+aligns to the standard it claims to, whether the task is developmentally plausible for the age on+the cover, and whether the alignment can be stated without reproducing text that is not yours.++That last constraint is not incidental. Most education standards are free to read and not free to+reproduce. Common Core is usable only with its required attribution notice; NGSS is copyrighted+with a registered trademark on its name. A generator that paraphrases a standard into a worksheet+header is a legal problem, not a formatting choice. The skills here treat citing the code and+reproducing the text as different acts, and the source catalog records which is permitted for each+publisher.+"""++[repo]+url = "https://github.com/cbrock84/headcount-education"+upstream = "https://github.com/cbrock84/headcount"++[[department]]+name = "education"+title = "Education"+description = "Standards alignment, learning-materials design, assessment construction, and developmentally appropriate practice."+keywords = ["curriculum", "standards", "assessment", "pedagogy"]+category = "education"++[emit]+exclude_departments = []