AI for Education

Build a single-class-period AI literacy lesson plan that maps to UNESCO's AI Competency Framework (for students or teachers), to state-mandated AI literacy standards (where they exist), and to a specific grade band — with explicit attention to the four pillars districts adopting AI literacy in 2026 are converging on: how AI works (technical foundations), responsible use (ethics and bias), critical evaluation (verifying AI output, recognizing limitations), and human-AI collaboration (knowing when to use AI and when not to). Output is teacher-ready: a time-stamped lesson, a hands-on student activity that does not require students to enter PII into a chatbot, an exit ticket aligned to the named competency, and a teacher reflection prompt for the post-lesson PLC. Designed for the 134 state-bill wave, the US ED Supplemental Priority on AI in Education, and the surge in district AI-literacy graduation requirements (Boston Public Schools' September 2026 mandate is the visible leading edge).

Audit an existing assignment for AI-outsourceability, identify the core learning it is meant to measure, and produce a redesigned version that assumes AI is present and still requires genuine student thinking. The output is a revised assignment prompt, updated rubric criteria, and a teacher rationale memo explaining what changed and why — ready to share with students, families, and instructional coaches.

Generate a teacher-ready academic vocabulary set from a source text or unit topic — Tier 2 (high-utility cross-curricular) and Tier 3 (domain-specific) words selected with the Beck / McKeown / Kucan robust-vocabulary criteria, paired with student-friendly definitions, multiple usage exposures across modalities (read, hear, speak, write), Bloom-tagged practice items, and WIDA-banded scaffolds for English Learners. Output is a complete instructional packet: word list with selection rationale, student-facing word cards with definitions and examples, three to five distinct practice tasks per word that hit different cognitive levels, a formative assessment, and a re-exposure plan across the unit. Designed for the documented vocabulary instruction gap that AI vocabulary tools surface in 2026 — most generators produce a list and definitions but stop short of the robust-vocabulary, multiple-exposures, and WIDA-banded scaffolding that the meta-analytic research shows is necessary for retention and transfer.

Generate a balanced, standards-aligned set of assessment items — multiple choice, short answer, constructed response, performance task — with answer keys, scoring guidance, and misconception-tagged distractors. The output is calibrated to specified cognitive levels (Bloom's, Depth of Knowledge) and is ready to drop into a quiz, unit test, or LMS item bank.

Produce a single, coherent course-level (or classroom-level) AI use disclosure package that a teacher can paste into a syllabus, course site, LMS landing page, parent letter, and assignment-instruction templates — built around the now-converging "traffic-light" model (red / yellow / green) and a task-by-task AI permission grid that maps each major assignment in the course to a specific AI-use posture, attribution standard, and verification expectation. The skill is the classroom-level companion to the district-level AI Acceptable Use Policy (which is governance, out of repo scope) and the lesson-level AI Literacy Lesson Plan Generator (which is a single-period lesson on AI-as-content); it sits in between, at the place where the syllabus becomes a workable agreement among teacher, students, and families. The pedagogical move is to replace one-line "AI is allowed / not allowed" syllabus statements (which produce ambiguous mid-task disputes that consume teacher time and damage trust) with task-specific permissions, a stable attribution standard, and a verify-against-non-AI-source expectation that the class can navigate without case-by-case adjudication. Designed for the post-2026-NYC-traffic-light, post-Ohio-July-1-mandate landscape where parent and student demand for transparent, predictable AI rules has crossed from preference to expectation.

Generate a teacher-ready set of critical-AI-literacy inquiry prompts that students answer before, during, and after any AI-assisted task — built around the six critical-AI-literacy domains converging in the 2026 framework literature: output evaluation, bias awareness, thinking ownership, system understanding, ethical awareness, and strategic use. The skill is distinct from the AI Literacy Lesson Plan Generator (which produces a single-period lesson with one named competency) — this skill produces a reusable inquiry set the teacher embeds into rubrics, journal prompts, exit tickets, conferences, and discussion protocols across many tasks. The pedagogical move is to push students from passive AI use ("does this answer look right?") toward active critical engagement ("what does this tool do to people, knowledge, and power?") so the questions become automatic over time. Designed for the post-AI-Literacy-101 phase: schools that have moved past "AI exists and here are its parts" and need to operationalize critical engagement during everyday content-area work.

Map existing lesson plans, units, scope-and-sequence documents, or course outlines to an official standards framework (Common Core, NGSS, C3, NCTM, state-specific frameworks including TEKS / NYSED / VA SOL / FL B.E.S.T. / CA CCSS, IB, AP, AERO/international), produce a defensible alignment matrix, quantify gaps by coverage, depth, and cognitive level, and — when needed — cross-walk between two frameworks (e.g., Common Core ↔ TEKS, or state → AP). The output is strong enough to bring to an accreditation review, curriculum audit, or school-board submission without further rework.

Turn a lesson plan or learning objective into a concrete, teacher-usable differentiation plan grounded in UDL, MTSS/RTI tiering, and Carol Ann Tomlinson's content/process/product/learning-environment/affect framework. The output tells the teacher what to change, for which students, at which point in the lesson — not a list of abstract strategies. It integrates with `lesson-plan-builder` (upstream) and `text-level-adjuster` (for passage-level reading adjustments).

Take a batch of evidence artifacts from a single student (writing samples, project deliverables, presentation slide decks, oral-defense notes, collaboration self-assessments, peer feedback, exit tickets, teacher observation notes, AI-tutor session logs, conferring records) gathered across a reporting window, and produce an evidence-anchored synthesis of the student's progress on the durable skills the district has named in its Portrait of a Graduate or equivalent framework — typically some subset of critical thinking, creative problem-solving, communication, collaboration, adaptability, personal responsibility, civic engagement, and global awareness. The output is a rubric-anchored evidence brief, organized one section per durable skill, with each section grounded in dated artifact references and rated against the district's durable-skills rubric, ready for use in a portfolio defense, a student-led conference, a parent conference, or a Portrait of a Graduate progress report.

Produce a short, objective-aligned exit ticket (typically 3–5 items) that surfaces whether students actually met the lesson's learning target in the final few minutes of class. The output is ready to drop into Google Forms / Microsoft Forms / Schoology / Canvas / Seesaw / Pear Deck / Nearpod / Mentimeter / Kahoot / a paper slip with no reformatting, includes a teacher-facing answer key with misconception-to-reteach mapping, and produces a next-day grouping recommendation calibrated to the likely response distribution. The ticket is designed for the final 3–5 minutes of class so it reads in under 60 seconds per item and produces actionable data the teacher can use that night for tomorrow's reteach.

Generate a reusable library of personalized-but-efficient feedback comments organized by skill, standard, performance level, and delivery channel — ready to copy into report cards, LMS gradebooks, progress reports, and family emails. Unlike per-student one-off comment drafting, a comment bank is built once per subject and grade band, then drawn from and lightly edited across the full reporting cycle. The result is consistent, criterion-referenced feedback at scale without sacrificing teacher voice.

Produce a fully-structured, teach-ready lesson plan from a topic and learning objective — with standards alignment, time-stamped activity sequence, built-in differentiation and formative checks, and a materials list. The output is designed to drop into a district lesson-plan template, a sub folder, or a coaching review with zero extra formatting.

Produce a classroom-ready rubric — analytic or holistic, single-point or multi-level — with observable criteria, calibrated performance descriptors, and standards alignment. The output can be handed to students as a success criteria document and used by teachers for consistent, defensible grading.

Construct a reusable system-prompt scaffold for a Socratic AI tutor agent — a tutor that asks rather than answers, scaffolds reasoning instead of supplying it, and adapts question depth to the student's response. Output is a deployable system prompt for the school's approved AI tool (Khanmigo Custom, MagicSchool Student-side, SchoolAI Spaces, district-licensed Copilot/Gemini for Education, ChatGPT Edu, or a custom GPT/Project), plus the teacher-facing companion: a scope statement, the off-topic and unsafe-topic stop conditions, the give-the-answer escape hatch, and a sample 4-turn dialogue showing what good and bad responses look like. Designed around the four design principles the Socratic-AI-tutor research literature has converged on: structured questioning patterns, adaptive scaffolding, reflective pauses, and human teacher oversight of transcripts.

Produce personalized, formative feedback on student work that is anchored to the assignment rubric, structured around Hattie & Timperley's feed-up / feed-back / feed-forward model, calibrated to the student's grade band, the stakes of the task, and the delivery channel (written comment, LMS, writing conference, report-card narrative), and safe to hand to a student without a revision pass by the teacher.

Produce a complete substitute-teacher plan for one class period, a full day, or a multi-day absence — including a header block with every emergency contact and procedure the sub will need to find on the first page, an attendance and entry routine, a self-contained instructional sequence the sub can run without the teacher's curriculum software, an independent student task that holds up if the sub is unfamiliar with the content, classroom-management scripts the sub reads aloud verbatim, accommodations-continuity for IEP/504/EL students named without surfacing protected-class status, an emergency quick-card for fire / lockdown / medical / student-crisis / weapon-threat events, and a structured note-back form the teacher can read in under two minutes on return. Output is explicitly written for a substitute who does not know the students, the classroom norms, or the unit, and who may have arrived ten minutes before the bell.

Audit which students the teacher has actually given 1:1 or small-group attention to over a recent window — across conferring, AI-tutor hint reviews, intervention check-ins, and live Q&A — and surface the students who have been under-visited, with a fairness-frame the teacher names explicitly and a next-session re-allocation plan. Output is a per-class, not per-student, artifact: a one-page attention-equity summary the teacher can act on in the next instructional block. Distinct from the existing PLC Agenda & Data Protocol Builder (team-level data conversation across a grade or department), the IEP Goal Progress Tracker (longitudinal goal data on identified students), and the Writing Conference Prep Generator (the next conferring conversation for one student): this skill is a routine in-classroom equity checkpoint that runs across the whole roster and is meant to be re-run weekly or biweekly.

Rewrite a passage of text up or down to a target reading level — specified as a Lexile band, grade band, or **WIDA proficiency band** (Entering / Emerging / Developing / Expanding / Bridging / Reaching) — while preserving every factual claim, key concept, central argument, and standards-anchor from the source. The output is a ready-to-distribute, differentiated version of the same passage with student-facing scaffolds (bold tier-3 vocabulary, inline definitions where the level requires it, cognate bridges where the home-language inventory supports them) and a teacher-facing readability note. Designed to integrate with the rest of the repo — the WIDA banding mirrors `differentiation-planner` and `student-feedback-generator`; pairs with `academic-vocabulary-builder` when the passage is part of a vocabulary-rich unit.

Generate a 5–7 minute writing-conference plan for a single student-and-draft pair following the workshop-model "research–decide–teach" structure: read the draft, identify what the writer is doing well, choose one teaching point that will transfer beyond this draft, name a mentor-text exemplar that shows the move, and produce a short conferring script the teacher can follow without taking the pen out of the student's hand. Output is a per-conference, not per-class, artifact — designed for the elementary writing-workshop classroom, the secondary writer's-workshop or studio classroom, the AP Lang / postsecondary composition revision conference, the ELL writing-conference rotation, and the IEP writing-goal individual conference. Distinct from the existing Student Feedback Generator (which produces written feedback on the artifact for the student to read after the fact) and the Rubric Generator (which produces the scoring instrument): this skill produces the live-talk-ready facilitation plan for the conferring conversation itself.

Turn a teacher's quick bullet-point notes into a polished, family-friendly recurring classroom newsletter — calibrated to the teacher's audience (single-classroom homeroom, specials, multi-teacher elementary team, secondary content-area, co-teaching pair), the platform it ships on (email, ParentSquare, ClassDojo Class Story, Bloomz, Smore, Remind announcement, Seesaw class blog, SIS announcement), the school's photo / image-release policy, the home-language mix in the room, and the district's mass-communication guidelines. Sections are consistent, the tone is warm, families get clear ways to support learning at home, and the artifact is built to be skim-readable on a phone in 60 seconds — without the teacher spending 45 minutes formatting on a Sunday night.

Draft a short, parent/guardian-facing message for one of the five recurring communication types a classroom teacher or specialist sends during a typical week — (1) progress update, (2) behavioral or academic concern, (3) event/logistics announcement, (4) absence/tardy/missing-work follow-up, (5) positive recognition / "good news" — calibrated for the recipient's relationship to the student (primary parent / shared-custody co-parent / non-custodial-with-rights / foster / kinship / agency case worker), the platform the family uses (email / ParentSquare / ClassDojo / Remind / Bloomz / SIS messenger), the home language and reading level on file, FERPA constraints on what the teacher may share, and the district's escalation rules for content that belongs in a formal channel rather than a teacher email. The output is short enough that a busy family will actually read it, fact-first enough to open conversation rather than close it, and routed correctly so the teacher does not accidentally create a legal record where one is not appropriate.

Produce a single conference-day preparation packet for one student-and-family conference: a one-page data summary, three to four prioritized talking points (strengths first, then growth areas, then partnership asks), a conference-day script for opening and closing, anticipated parent concerns with prepared responses, and a structured follow-up plan with a default 10-day check-in. Output is teacher-facing, fits on the desk during a 15–20 minute conference window, and is built to be skim-readable in 90 seconds before the family walks in. Designed for the high-velocity reality of conference week — most elementary teachers run 20–25 conferences in 2 days; secondary teachers run 50+ in a half-day cycle.

Draft a structured behavior intervention plan (BIP) or positive behavior support plan from a teacher's description of a target behavior, suspected function, and classroom context. The output is a draft — not a final plan — designed to jump-start an IEP team or MTSS conversation, calibrated to the tier of support the data justifies (Tier 2 classroom-level / Tier 3 IEP-anchored), with named district-policy hooks (PBIS framework, restraint/seclusion policy reference, crisis-team contact, school-psych or BCBA referral pathway, parent-input requirement) pulled from config so the plan integrates with the building's existing systems rather than living as a standalone document. The plan never substitutes for the FBA/BIP process required under IDEA when a student's behavior impedes learning, and never recommends physical restraint or seclusion — those route to the building's trained crisis team under district policy.

Turn raw classroom-walkthrough notes or a lesson transcript into a framework-aligned, coaching-oriented feedback report for a teacher. The output is written for the instructional coach, principal, assistant principal, or department chair — not for the teacher directly — and packages evidence, alignment to the district's observation framework (Danielson, Marzano, NIET TAP, state model, or custom), a small number of high-leverage growth moves, and a pre-filled post-observation conference agenda. The feedback is explicitly coaching-first and positions the observer as a thinking partner rather than an evaluator.

Draft a competitive, funder-aligned education grant proposal — Letter of Inquiry (LOI), full narrative, or supplemental request — that maps the applicant's program, evidence base, and budget directly to the named priorities of the funding source. Output is structured for grant-writer review and program-officer readability, with priority-language alignment on every section, an evidence-of-need block grounded in the applicant's own data, and a logic-model + measurement plan a program officer can score in under five minutes. Designed for the wave of 2026 federal and private AI-in-education funding (US ED Supplemental Priority effective 2026-05-13, NSF K-12 AI supplements up to $300K, Digital Promise K-12 AI Infrastructure Program $50K–$250K, Humanity AI $500M five-year pooled fund, FIPSE postsecondary, Spencer Foundation AI + Education) — but the structure works for any standards-based education grant.

Convert raw progress-monitoring data (work samples, curriculum-based measures, teacher observation logs, assistive tech usage, behavior frequency counts) into a legally-defensible narrative progress report for an annual, quarterly, or mid-year IEP review. The output is case-manager-ready: it ties each goal to its measurable criterion, reports current performance in the same units as the baseline, and states whether the student is on track, making progress but not at the projected rate, or not making progress — which triggers a team discussion under IDEA.

Produce an **advisory-only** menu of accommodation and modification candidates for a single student profile — for case-manager review, IEP-team discussion, or 504-plan drafting — drawn from the categories used in IDEA accommodation guidance, Section 504 of the Rehabilitation Act, the WIDA Accessibility & Accommodations framework, and the major state assessment accommodation manuals. The output is **never** a finalized accommodation list, never enters the IEP or 504 document verbatim, and is never shared with families before the team has reviewed it. Every recommended accommodation is paired with the source category it draws from, the access barrier it is meant to address, the evidence base (or absence of evidence), and a "team-discussion question" the case manager would raise at the IEP/504 meeting before the accommodation is adopted.

Build a meeting-ready agenda and paired data-analysis protocol for a professional learning community, grade-level team, or department PLC — grounded in the DuFour four-question framework (What do we want students to learn? How will we know they learned it? What will we do if they didn't? What will we do if they did?) and designed so the meeting produces concrete next-instructional-move commitments, not discussion-for-discussion's-sake. Output is structured for the team facilitator, with explicit roles, a time-stamped agenda, a protocol matched to the data on the table, and a commitment-tracking block.

Draft a facilitator script for a restorative conversation, classroom circle, or formal restorative conference following a harm incident — for use by a trained teacher, dean, counselor, or restorative-practices coordinator. The output includes the facilitator's preamble, the sequenced restorative questions for the person(s) responsible, the person(s) harmed, and supporters, an agreement-to-repair template, and follow-up checkpoints. It is explicitly scoped as a **draft for a trained facilitator** — not a substitute for certified training in restorative practice or a replacement for a formal mediation/Title IX/CPS process where one is required.

Turn raw student performance data (grades, standards scores, attendance, assignment completion, behavior and work-habits observations) into polished narrative progress reports that (a) match the school's report-card format (traditional letter-grade, standards-based, or narrative), (b) use the correct register for the grade band, (c) follow subject-specific narrative conventions, (d) are FERPA-compliant, and (e) flag students who need intervention under the school's MTSS/RTI framework.

Turn rough notes into a ready-to-send education email tailored to the recipient, the situation, and the school's communication norms — covering parent/guardian outreach, colleague coordination, administrator correspondence, vendor/partner communication, age-appropriate student messages, and high-sensitivity scenarios that have to be handled with explicit FERPA, custody, and language-access discipline. The output is a finished draft the user can paste into Gmail, Outlook, ParentSquare, ClassDojo, Remind, Bloomz, or the district SIS messenger with minimal editing — and a separate "do not send via email" flag when the topic must be routed to a formal channel instead.

Turn raw education-meeting notes into a structured summary that names decisions, assigns action items with owner-and-deadline, surfaces follow-ups, captures key dates, and applies a meeting-type-specific structure that respects the legal, FERPA, and team-process rules each kind of meeting carries. Output is paired with an attendee-distribution rule (what each role-group should and should not see), a separation between observation / interpretation / action notes, and an explicit "what is missing for compliance" flag where required components were not addressed.

Draft a professional, platform-appropriate, FERPA-clean response to a public review of a school, program, or educational organization — calibrated to the platform's audience and norms (Google Business Profile, GreatSchools, Niche, Facebook recommendations, Yelp, district parent-portal surveys, state report-card commentary, accreditation-survey commentary), to the review's sentiment, and to whether the content crosses into a legal, civil-rights, Title IX, CPS, or accreditation-impact line that requires admin/legal review before posting. Output is a paste-ready response plus an explicit triage flag for any review that should not be answered publicly until a named human has reviewed it.