A useful agent starts with a sound use case.
Choose a practice, examine its limits, and keep educational judgment in human hands. These are contextual editorial recommendations.
Under review. These entries are under review and have not yet passed all eight gates of Methodology 1.0.
Suggest a use case
Under review inappropriateAgent determines student eligibility, services or placement from records.
Human in the loop: None. This is the failure. Limits: Uses protected learner records and makes consequential decisions about students. Methodology 1.0: Publish as warning only.
Under review conditionalReads a unit outline the teacher provides, drafts practice questions with answer keys tied to source passages, and flags questions it cannot ground in the material.
Human in the loop: Teacher reviews, edits and selects every question before learners see it. Limits: Questions can be wrong or off-level. Agent must not grade or answer for students. Methodology 1.0: Hold.
How to run: Draft practice question sets from teacher-supplied materials
- Reads a unit outline the teacher provides, drafts practice questions with answer keys tied to source passages, and flags questions it cannot ground in the material.
- Teacher reviews, edits and selects every question before learners see it.
- Questions can be wrong or off-level. Agent must not grade or answer for students.
Sources: MagicSchool AI platform for school districts ↗Under review conditionalTakes a teacher lesson plan and proposes alternative means of representation, engagement and expression mapped to UDL guideline checkpoints.
Human in the loop: Teacher decides which options to use and adapts them to the class. Limits: Suggestions are generic; no student profiles are used. Not an accommodation decision. Methodology 1.0: Hold.
How to run: Differentiate a lesson using UDL options
- Takes a teacher lesson plan and proposes alternative means of representation, engagement and expression mapped to UDL guideline checkpoints.
- Teacher decides which options to use and adapts them to the class.
- Suggestions are generic; no student profiles are used. Not an accommodation decision.
Sources: CAST UDL Guidelines ↗MagicSchool AI platform for school districts ↗Under review conditionalDrafts an analytic rubric from the teacher objective and task prompt, then checks each criterion against the objective and lists gaps.
Human in the loop: Teacher finalizes criteria and performance levels. Limits: Rubric quality depends on the objective. Agent does not score student work. Methodology 1.0: Hold.
How to run: Build a rubric draft aligned to a stated objective
- Drafts an analytic rubric from the teacher objective and task prompt, then checks each criterion against the objective and lists gaps.
- Teacher finalizes criteria and performance levels.
- Rubric quality depends on the objective. Agent does not score student work.
Sources: ISTE+ASCD: AI in Education ↗Under review conditionalTranslates a teacher-written family letter into requested languages and produces a back-translation for spot checks.
Human in the loop: Teacher or a fluent staff member reviews before sending. Limits: Machine translation errors; no student names or records in the input. Methodology 1.0: Hold.
Under review conditionalSearches the What Works Clearinghouse for a named practice, summarizes the review ratings with links, and states when no review exists.
Human in the loop: Teacher reads the cited reviews before acting. Limits: Summaries can overstate findings; evidence may not match the teacher context. Methodology 1.0: Hold.
How to run: Find vetted research on an instructional strategy
- Searches the What Works Clearinghouse for a named practice, summarizes the review ratings with links, and states when no review exists.
- Teacher reads the cited reviews before acting.
- Summaries can overstate findings; evidence may not match the teacher context.
Sources: IES What Works Clearinghouse ↗Under review conditionalCompares the school draft against published state AI guidance and produces a gap table with citations.
Human in the loop: Principal and staff committee decide policy language. Limits: Not legal advice. State guidance changes; links must be rechecked. Methodology 1.0: Hold.
How to run: Draft a staff AI-use guidance page from state guidance
- Compares the school draft against published state AI guidance and produces a gap table with citations.
- Principal and staff committee decide policy language.
- Not legal advice. State guidance changes; links must be rechecked.
Sources: AI for Education: State AI Guidance for Education ↗Under review conditionalPulls the leader's state entry from a public tracker, summarizes key expectations with links, and lists open questions.
Human in the loop: Leader verifies against the official state document. Limits: Tracker can lag official releases. Methodology 1.0: Hold.
Under review conditionalGathers the vendor public privacy and product pages and drafts answers to a review checklist, marking each answer sourced or unknown.
Human in the loop: Leader and IT make the approval decision. Limits: Vendor claims are self-reported; agent does not approve tools. Methodology 1.0: Hold.
Under review conditionalTurns a leader list of topics into a timed agenda and a notes template with action-item fields.
Human in the loop: Leader edits and owns the agenda. Limits: Low risk; a simple template may do the job as well (Gate 1 comparison needed). Methodology 1.0: Hold.
Under review conditionalSearches Grants.gov for listings matching district criteria, summarizes eligibility and deadlines with links, and flags uncertain fits.
Human in the loop: Grants office decides which to pursue. Limits: Listings change; agent does not submit applications. Methodology 1.0: Hold.
How to run: Screen federal grant opportunities
- Searches Grants.gov for listings matching district criteria, summarizes eligibility and deadlines with links, and flags uncertain fits.
- Grants office decides which to pursue.
- Listings change; agent does not submit applications.
Sources: Grants.gov ↗Under review conditionalReads USAC public E-Rate program pages and drafts an internal calendar of filing windows and steps with links.
Human in the loop: E-Rate coordinator verifies dates before use. Limits: Dates must be confirmed with USAC each cycle. No filings by the agent. Methodology 1.0: Hold.
How to run: Draft E-Rate filing calendar reminders
- Reads USAC public E-Rate program pages and drafts an internal calendar of filing windows and steps with links.
- E-Rate coordinator verifies dates before use.
- Dates must be confirmed with USAC each cycle. No filings by the agent.
Sources: USAC: E-Rate ↗Under review conditionalMonitors USDA FNS program pages, drafts a plain summary of changes with links for nutrition staff.
Human in the loop: Nutrition director reviews before sharing. Limits: Summaries are not compliance determinations. Methodology 1.0: Hold.
How to run: Summarize school meal program guidance updates
- Monitors USDA FNS program pages, drafts a plain summary of changes with links for nutrition staff.
- Nutrition director reviews before sharing.
- Summaries are not compliance determinations.
Sources: USDA FNS: National School Lunch Program ↗Under review conditionalReads public board policy documents the district supplies and drafts an FAQ with citations to policy sections.
Human in the loop: Policy office approves the FAQ before posting. Limits: Must quote policy accurately; not legal interpretation. Methodology 1.0: Hold.
Under review conditionalClassifies incoming staff tickets by category and urgency, suggests a routing queue and a draft reply.
Human in the loop: Technician approves routing changes and replies. Limits: Misclassification; tickets may contain sensitive data, so scope must exclude student records. Methodology 1.0: Hold.
Under review conditionalMaps the district's public-facing practices against CISA K-12 recommendations and drafts a gap checklist.
Human in the loop: IT director validates findings and priorities. Limits: Not an audit; agent has no system access. Methodology 1.0: Hold.
Under review conditionalDrafts a crosswalk of existing written IT procedures to NIST CSF functions and lists unmapped areas.
Human in the loop: Security lead reviews and owns the crosswalk. Limits: Documentation-only; does not verify controls work. Methodology 1.0: Hold.
How to run: Map security controls to the NIST CSF
- Drafts a crosswalk of existing written IT procedures to NIST CSF functions and lists unmapped areas.
- Security lead reviews and owns the crosswalk.
- Documentation-only; does not verify controls work.
Sources: NIST Cybersecurity Framework ↗Under review conditionalConfigures a retrieval agent over published IT knowledge base articles that answers staff questions with article links and escalates to a ticket when unsure.
Human in the loop: IT owns the knowledge base and reviews escalations. Limits: Stale articles; must not take account actions. Methodology 1.0: Hold.
Under review conditionalDrafts posts from approved district announcements and places them in a scheduling queue as drafts.
Human in the loop: Communications staff approve every post before it publishes. Limits: No student photos or names without consent. Tone and accuracy errors. Methodology 1.0: Hold.
How to run: Draft and schedule social media posts
- Drafts posts from approved district announcements and places them in a scheduling queue as drafts.
- Communications staff approve every post before it publishes.
- No student photos or names without consent. Tone and accuracy errors.
Sources: Hootsuite social media scheduler ↗n8n Docs: Integrate AI ↗Under review conditionalRuns automated checks on public pages, lists WCAG failures with page links, and drafts fix tickets.
Human in the loop: Web staff verify and fix; manual testing is still required. Limits: Automated tools miss many issues. Methodology 1.0: Hold.
Under review conditionalCollects approved announcements and drafts a newsletter with consistent formatting and plain language.
Human in the loop: Communications lead approves before sending. Limits: Agent must not send; check for errors and dates. Methodology 1.0: Hold.
Under review conditionalProposes alt text for public non-student images and flags images that need human description.
Human in the loop: Web editor approves alt text. Limits: Context errors; images of people need human handling. Methodology 1.0: Hold.
Under review conditionalDrafts plain-language explainers from public parent center resources, with links.
Human in the loop: Special education director approves content. Limits: No student data. Not legal advice about rights. Methodology 1.0: Hold.
How to run: Draft family-facing explainers of the special education process
- Drafts plain-language explainers from public parent center resources, with links.
- Special education director approves content.
- No student data. Not legal advice about rights.
Sources: Center for Parent Information and Resources ↗Under review conditionalCurates and summarizes public resources on UDL and inclusive practice into a staff resource list with links.
Human in the loop: Coordinator reviews the list. Limits: Resource quality varies. Methodology 1.0: Hold.
Under review conditionalReads an instructor syllabus and drafts an FAQ page with section citations, flagging gaps.
Human in the loop: Instructor approves before posting in the LMS. Limits: Must not invent policies; students still need the syllabus. Methodology 1.0: Hold.
Under review conditionalAnswers prospective student questions from public institutional web pages with links, and routes anything account-specific to staff.
Human in the loop: Staff own answers to anything involving records or decisions. Limits: Hallucination risk; no access to student records. Methodology 1.0: Hold.
Under review conditionalCompares public guidance and drafts an options memo with citations.
Human in the loop: Faculty senate decides policy. Limits: Not a recommendation of a single policy. Methodology 1.0: Hold.
How to run: Summarize institutional AI policy options for a faculty senate
- Compares public guidance and drafts an options memo with citations.
- Faculty senate decides policy.
- Not a recommendation of a single policy.
Sources: UNESCO: AI in education ↗NIST AI Risk Management Framework ↗Digital Promise: AI in Education (source link removed: the page moved and the original document could not be verified)Under review conditionalReads a public funding notice and drafts a requirements checklist with section references.
Human in the loop: Research office verifies against the notice. Limits: Notices change; agent does not submit. Methodology 1.0: Hold.
How to run: Draft research grant boilerplate checklist
- Reads a public funding notice and drafts a requirements checklist with section references.
- Research office verifies against the notice.
- Notices change; agent does not submit.
Sources: Grants.gov ↗Under review conditionalReads the school's or district's existing food allergy management plan (the general plan, not any individual child's plan) and drafts a gap checklist against the five priority areas CDC says each Food Allergy Management Prevention Plan should address: daily management, emergency preparedness, staff professional development, education for children and families, and a safe environment.
Human in the loop: School nurse or health services lead reviews every gap and decides any plan change. Limits: Not medical or legal advice; state guidelines differ (CDC links to state guidelines). No individual health plans, health records or student names as inputs. Methodology 1.0: Hold.
How to run: Check a school food allergy plan against CDC priority areas
- Reads the school's or district's existing food allergy management plan (the general plan, not any individual child's plan) and drafts a gap checklist against the five priority areas CDC says each Food Allergy Management Prevention Plan should address: daily management, emergency preparedness, staff professional development, education for children and families, and a safe environment.
- School nurse or health services lead reviews every gap and decides any plan change.
- Not medical or legal advice; state guidelines differ (CDC links to state guidelines). No individual health plans, health records or student names as inputs.
Sources: Food Allergies in Schools | Managing Health Conditions in School | CDC ↗ (Snapshot pending)Under review conditionalDrafts an agenda and planning checklist for a local Head Start and elementary school transition summit, organized around the four points of connection HeadStart.gov names (family-school, child-school, program-school, community-school) and linking the HeadStart.gov summit guide, slides and planning tools.
Human in the loop: Head Start director and elementary principal approve the agenda and invitations. Limits: Planning only; no child assessment data, IEP information or family names as inputs. Methodology 1.0: Hold.
Under review conditionalDrafts a building walkthrough plan and task list for a school indoor air quality (IAQ) team, drawn from EPA's IAQ Tools for Schools resources (the Framework for Effective School IAQ Management, the Action Kit and the Preventive Maintenance Guidance), covering the areas EPA's walkthrough checklists address: ventilation, cleaning and maintenance, environmental asthma triggers, radon and integrated pest management.
Human in the loop: Facilities director and the school IAQ coordinator approve the plan and assign every walkthrough task. Limits: Planning only; does not replace testing, a licensed professional, or state and local building requirements. No student or staff health records as inputs. Methodology 1.0: Hold.
How to run: Draft an indoor air quality walkthrough plan from EPA guidance
- Drafts a building walkthrough plan and task list for a school indoor air quality (IAQ) team, drawn from EPA's IAQ Tools for Schools resources (the Framework for Effective School IAQ Management, the Action Kit and the Preventive Maintenance Guidance), covering the areas EPA's walkthrough checklists address: ventilation, cleaning and maintenance, environmental asthma triggers, radon and integrated pest management.
- Facilities director and the school IAQ coordinator approve the plan and assign every walkthrough task.
- Planning only; does not replace testing, a licensed professional, or state and local building requirements. No student or staff health records as inputs.
Sources: IAQ Tools for Schools Resources | US EPA ↗ (Snapshot pending)Indoor Air Quality Tools for Schools | US EPA ↗ (Snapshot pending)Under review conditionalTakes a teacher's list of openly licensed images, texts and music used in a slide deck or handout and drafts an attribution line for each one using the Creative Commons TASL elements (Title, Author, Source, License), flags items whose license details are missing, and flags any planned change to an ND (NoDerivatives) licensed work, since CC says those licenses do not allow adaptations.
Human in the loop: Teacher or school librarian checks every line against the original source page before the material is shared. Limits: CC says there is no single correct form of attribution and it must reasonably include what the licensor supplied, so the agent drafts and flags, it does not certify compliance. Not legal advice. CC also advises against putting attribution in alt text, so the agent keeps credits out of alt text. No student work or names as inputs. Methodology 1.0: Hold.
How to run: Draft Creative Commons attribution lines for openly licensed classroom materials
- Takes a teacher's list of openly licensed images, texts and music used in a slide deck or handout and drafts an attribution line for each one using the Creative Commons TASL elements (Title, Author, Source, License), flags items whose license details are missing, and flags any planned change to an ND (NoDerivatives) licensed work, since CC says those licenses do not allow adaptations.
- Teacher or school librarian checks every line against the original source page before the material is shared.
- CC says there is no single correct form of attribution and it must reasonably include what the licensor supplied, so the agent drafts and flags, it does not certify compliance. Not legal advice. CC also advises against putting attribution in alt text, so the agent keeps credits out of alt text. No student work or names as inputs.
Sources: Recommended practices for attribution | Creative Commons ↗ (Snapshot pending)Under review conditionalReads a district or state CTE team's written work-based learning (WBL) program description and drafts a gap sheet against the three components the U.S. Department of Education WBL Toolkit names for comprehensive programs (alignment of classroom and workplace learning; application of academic, technical and employability skills in a work setting; support from classroom or workplace mentors), plus the toolkit's sections on engaging employers, collecting data and scaling programs, with questions for the team.
Human in the loop: CTE director reviews the sheet with program staff and employer partners and decides any program change. Limits: The agent drafts questions, not a quality rating. Perkins V, ESSA and WIOA requirements, which the toolkit references, are confirmed by staff against the law and state guidance. No student placement, wage or participation records as inputs. Methodology 1.0: Hold.
How to run: Draft a work-based learning program gap check from the federal WBL toolkit
- Reads a district or state CTE team's written work-based learning (WBL) program description and drafts a gap sheet against the three components the U.S. Department of Education WBL Toolkit names for comprehensive programs (alignment of classroom and workplace learning; application of academic, technical and employability skills in a work setting; support from classroom or workplace mentors), plus the toolkit's sections on engaging employers, collecting data and scaling programs, with questions for the team.
- CTE director reviews the sheet with program staff and employer partners and decides any program change.
- The agent drafts questions, not a quality rating. Perkins V, ESSA and WIOA requirements, which the toolkit references, are confirmed by staff against the law and state guidance. No student placement, wage or participation records as inputs.
Sources: Work-Based Learning Toolkit | U.S. Department of Education, Perkins Collaborative Resource Network ↗ (Snapshot pending)Under review conditionalReads EPA's public Clean School Bus (CSB) program page on a set schedule and drafts a one-page status brief for a district transportation director: what the page currently says about open or upcoming funding, the dates it names, and what changed since the last brief, with links to the rebate documents and awards pages it lists.
Human in the loop: Transportation director and business office review the brief before any funding decision or board item. Limits: Program rules are changing, so the brief quotes the page and dates it, and never states eligibility or award odds. Any application step is confirmed by staff with EPA directly. No student or route data as inputs. Methodology 1.0: Hold.
How to run: Draft a Clean School Bus program status brief for a transportation director
- Reads EPA's public Clean School Bus (CSB) program page on a set schedule and drafts a one-page status brief for a district transportation director: what the page currently says about open or upcoming funding, the dates it names, and what changed since the last brief, with links to the rebate documents and awards pages it lists.
- Transportation director and business office review the brief before any funding decision or board item.
- Program rules are changing, so the brief quotes the page and dates it, and never states eligibility or award odds. Any application step is confirmed by staff with EPA directly. No student or route data as inputs.
Sources: Clean School Bus Program | US EPA ↗ (Snapshot pending)Under review conditionalTakes a college's published Satisfactory Academic Progress (SAP) policy text (public, no student data) and drafts a gap list for the financial aid office against the 2025-2026 FSA Handbook, Volume 1, Chapter 1: whether the policy states both qualitative and quantitative standards, whether it is at least as strict as the policy for students not receiving Title IV aid in the same program, how and when progress is checked, whether rounding rules are stated if used, and whether financial aid warning, appeal, probation and academic plan terms match how the handbook describes them (for example, warning is only for schools that check SAP every payment period and lasts one payment period). Each gap cites the handbook section it comes from.
Human in the loop: Financial aid director reviews every gap and decides any policy change; the workflow never evaluates an individual student's progress or appeal. Limits: The handbook's worked examples (such as a 2/3 pace and 150 percent maximum timeframe) illustrate one school's policy and must not be presented as requirements. Federal rules change by award year, so the gap list names the handbook year used. Not legal advice. Individual SAP determinations and appeals are consequential decisions about students and stay with staff (see the aed-uc-149 warning). No student records as inputs. Methodology 1.0: Hold.
How to run: Review a college's Satisfactory Academic Progress policy against the FSA Handbook
- Takes a college's published Satisfactory Academic Progress (SAP) policy text (public, no student data) and drafts a gap list for the financial aid office against the 2025-2026 FSA Handbook, Volume 1, Chapter 1: whether the policy states both qualitative and quantitative standards, whether it is at least as strict as the policy for students not receiving Title IV aid in the same program, how and when progress is checked, whether rounding rules are stated if used, and whether financial aid warning, appeal, probation and academic plan terms match how the handbook describes them (for example, warning is only for schools that check SAP every payment period and lasts one payment period). Each gap cites the handbook section it comes from.
- Financial aid director reviews every gap and decides any policy change; the workflow never evaluates an individual student's progress or appeal.
- The handbook's worked examples (such as a 2/3 pace and 150 percent maximum timeframe) illustrate one school's policy and must not be presented as requirements. Federal rules change by award year, so the gap list names the handbook year used. Not legal advice. Individual SAP determinations and appeals are consequential decisions about students and stay with staff (see the aed-uc-149 warning). No student records as inputs.
Sources: School-Determined Requirements | 2025-2026 Federal Student Aid Handbook ↗ (Snapshot pending)Under review conditionalTakes a list of course names and the local retention rule (no student data) and drafts a per-course reset checklist from MoodleDocs "Reset course": confirm a backup including user data exists first (MoodleDocs says reset is irreversible), choose General, Role, Gradebook, Group and Activity options, set a new course start date (all dates shift by the same amount), and note that course logs are not deleted and grades remain recorded against user accounts. Adds the Moodle 5.2 asynchronous reset setting and where admins monitor it.
Human in the loop: LMS admin or teacher clicks Reset; the workflow never resets anything itself. Limits: Irreversible data deletion; local records retention rules may require keeping grades longer (UNKNOWN per district). Version-specific: asynchronous reset is described for Moodle 5.2. Methodology 1.0: Strong candidate for replay.
How to run: Draft a Moodle end-of-term course reset checklist with a backup-first gate
- Takes a list of course names and the local retention rule (no student data) and drafts a per-course reset checklist from MoodleDocs "Reset course": confirm a backup including user data exists first (MoodleDocs says reset is irreversible), choose General, Role, Gradebook, Group and Activity options, set a new course start date (all dates shift by the same amount), and note that course logs are not deleted and grades remain recorded against user accounts. Adds the Moodle 5.2 asynchronous reset setting and where admins monitor it.
- LMS admin or teacher clicks Reset; the workflow never resets anything itself.
- Irreversible data deletion; local records retention rules may require keeping grades longer (UNKNOWN per district). Version-specific: asynchronous reset is described for Moodle 5.2.
Sources: Reset course | MoodleDocs ↗ (Snapshot pending)Under review conditionalTakes the district's local SIS code tables (code values and labels only, for example attendance or exit codes; no student records) and drafts a gap list against Ed-Fi default descriptors: which local codes map cleanly to ed-fi.org namespace values, which have no match, and which are in locally defined namespaces. Names the Data Standard version used (Ed-Fi lists v6.1.0 as current, and notes v6.0 brought breaking API changes).
Human in the loop: Data coordinator decides every mapping; the state agency's own reporting manual governs (UNKNOWN which states use Ed-Fi and which descriptors they require). Limits: Wrong mappings change state-reported counts, a consequential outcome that stays with staff. PowerSchool, Infinite Campus and Skyward Ed-Fi connectors: UNKNOWN (not verified today). Methodology 1.0: Strong candidate for replay (with Grand Bend ISD fictitious sample data, which Ed-Fi publishes).
How to run: Draft a local-code to Ed-Fi descriptor mapping gap list for state reporting prep
- Takes the district's local SIS code tables (code values and labels only, for example attendance or exit codes; no student records) and drafts a gap list against Ed-Fi default descriptors: which local codes map cleanly to ed-fi.org namespace values, which have no match, and which are in locally defined namespaces. Names the Data Standard version used (Ed-Fi lists v6.1.0 as current, and notes v6.0 brought breaking API changes).
- Data coordinator decides every mapping; the state agency's own reporting manual governs (UNKNOWN which states use Ed-Fi and which descriptors they require).
- Wrong mappings change state-reported counts, a consequential outcome that stays with staff. PowerSchool, Infinite Campus and Skyward Ed-Fi connectors: UNKNOWN (not verified today).
Sources: Ed-Fi Data Standard | Ed-Fi Alliance ↗ (Snapshot pending)Under review conditionalTakes the business office's list of federal awards (program names, Assistance Listing numbers, pass-through entities and award IDs; no personal data) and drafts a SEFA readiness checklist from 2 CFR 200.510(b): programs listed by federal agency with Assistance Listing numbers, cluster names and totals, pass-through entity names and identifying numbers, totals to subrecipients, loan balances in notes, and a note on significant accounting policies and whether the de minimis indirect rate (up to 15 percent) was used. Adds the 200.501 threshold check ($1,000,000 or more in federal expenditures triggers a single or program-specific audit).
Human in the loop: Business manager and auditor own the SEFA; the workflow never computes final expenditure totals. Limits: Thresholds and rules change with Uniform Guidance revisions; the draft names the eCFR date used. Not audit or legal advice. Methodology 1.0: Strong candidate for replay.
How to run: Draft a Schedule of Expenditures of Federal Awards (SEFA) prep checklist from 2 CFR 200.510
- Takes the business office's list of federal awards (program names, Assistance Listing numbers, pass-through entities and award IDs; no personal data) and drafts a SEFA readiness checklist from 2 CFR 200.510(b): programs listed by federal agency with Assistance Listing numbers, cluster names and totals, pass-through entity names and identifying numbers, totals to subrecipients, loan balances in notes, and a note on significant accounting policies and whether the de minimis indirect rate (up to 15 percent) was used. Adds the 200.501 threshold check ($1,000,000 or more in federal expenditures triggers a single or program-specific audit).
- Business manager and auditor own the SEFA; the workflow never computes final expenditure totals.
- Thresholds and rules change with Uniform Guidance revisions; the draft names the eCFR date used. Not audit or legal advice.
Sources: 2 CFR 200.510 Financial statements | eCFR ↗ (Snapshot pending)2 CFR 200.501 Audit requirements | eCFR ↗ (Snapshot pending)Under review conditionalReads a school's or district's existing family engagement plan and drafts a reflection sheet mapping it to the sections of the Dual Capacity-Building Framework for Family-School Partnerships (Version 2): the challenge, essential conditions, policy and program goals, and capacity outcomes, with questions for the planning team on gaps.
Human in the loop: Family engagement lead and principal review the sheet with families and staff and decide any plan change. Limits: The framework's authors say it is a compass, not a blueprint, and must be fit to local context; the agent drafts questions, not a score. Framework is copyrighted (Mapp and Bergman, 2019), so cite it and do not reproduce it. No family names or contact data as inputs. Methodology 1.0: Hold.
How to run: Review a family engagement plan against the Dual Capacity Framework
- Reads a school's or district's existing family engagement plan and drafts a reflection sheet mapping it to the sections of the Dual Capacity-Building Framework for Family-School Partnerships (Version 2): the challenge, essential conditions, policy and program goals, and capacity outcomes, with questions for the planning team on gaps.
- Family engagement lead and principal review the sheet with families and staff and decide any plan change.
- The framework's authors say it is a compass, not a blueprint, and must be fit to local context; the agent drafts questions, not a score. Framework is copyrighted (Mapp and Bergman, 2019), so cite it and do not reproduce it. No family names or contact data as inputs.
Sources: Dual Capacity-Building Framework for Family-School Partnerships ↗ (Snapshot pending)Under review conditionalTakes a program's planned staff training calendar (session titles, dates and target roles only; no staff names, records or child information) and drafts a coverage matrix against 45 CFR 1302.47(b)(4). Staff with regular child contact need initial orientation within three months of hire and ongoing training that covers, as appropriate to role and child ages: infectious disease prevention; SIDS prevention and safe sleep; medication administration with parental consent; food and allergic reaction emergencies; building and premises safety, including water and vehicle hazards; prevention of shaken baby syndrome, abusive head trauma and maltreatment; emergency preparedness; hazardous materials and biocontaminants; transportation precautions if applicable; first aid and CPR; and recognizing and reporting child abuse and neglect. Staff without regular child contact need orientation within three months, training on the health and safety rules for their work, and emergency and disaster procedures. Each topic is marked covered, not found, or needs review, with the calendar session cited.
Human in the loop: The health and safety manager confirms every match and adds state, local and tribal requirements; the workflow does not approve the training plan or judge any staff member. Limits: State licensing and tribal rules add topics and hours the manager must supply. The regulation was amended as recently as Dec 19, 2025 (90 FR 59400 per the fetched page), so the current text must be rechecked before use. Not legal advice. Methodology 1.0: Strong candidate for replay.
How to run: Map a Head Start program's staff safety training calendar to the required training topics
- Takes a program's planned staff training calendar (session titles, dates and target roles only; no staff names, records or child information) and drafts a coverage matrix against 45 CFR 1302.47(b)(4). Staff with regular child contact need initial orientation within three months of hire and ongoing training that covers, as appropriate to role and child ages: infectious disease prevention; SIDS prevention and safe sleep; medication administration with parental consent; food and allergic reaction emergencies; building and premises safety, including water and vehicle hazards; prevention of shaken baby syndrome, abusive head trauma and maltreatment; emergency preparedness; hazardous materials and biocontaminants; transportation precautions if applicable; first aid and CPR; and recognizing and reporting child abuse and neglect. Staff without regular child contact need orientation within three months, training on the health and safety rules for their work, and emergency and disaster procedures. Each topic is marked covered, not found, or needs review, with the calendar session cited.
- The health and safety manager confirms every match and adds state, local and tribal requirements; the workflow does not approve the training plan or judge any staff member.
- State licensing and tribal rules add topics and hours the manager must supply. The regulation was amended as recently as Dec 19, 2025 (90 FR 59400 per the fetched page), so the current text must be rechecked before use. Not legal advice.
Sources: 45 CFR 1302.47 Safety practices | Cornell LII ↗ (Snapshot pending)Under review conditionalTakes the district's onboarding calendar and process notes (who completes Section 2, whether the district uses E-Verify; no employee names or documents) and drafts an HR checklist from USCIS I-9 Central: Form I-9 is required for every person hired for pay; its parts are Section 1 (employee), Section 2 (employer or authorized representative), Supplement A (preparer or translator) and Supplement B (reverification and rehires); Section 2 must be completed within 3 business days of the first day of work for pay (for example, a Monday start means Thursday), or by the first day if the job lasts under 3 days; employees choose one List A document or one List B plus one List C document, and staff must not ask for a specific list; under E-Verify a List B document must have a photo and remote examination is only for E-Verify participants. Output is a timeline per start-date pattern and a checklist of employer duties (examine originals, record document details, enter the physical address and not a P.O. box, return documents).
Human in the loop: HR staff or the named authorized representative examines every document and signs Section 2; the workflow never sees, stores or judges an employee's documents. Limits: Not legal advice; minors, special placement and unfamiliar documents need USCIS Special Categories and the M-274 handbook, which are flagged rather than summarized. Telling an employee which documents to present is a compliance risk, so the checklist must keep the choice with the employee. Form versions change, so the checklist names the USCIS page review date (Sep 18, 2026). Methodology 1.0: Strong candidate for replay.
How to run: Draft a Form I-9 completion timeline and checklist for a district's new hire onboarding
- Takes the district's onboarding calendar and process notes (who completes Section 2, whether the district uses E-Verify; no employee names or documents) and drafts an HR checklist from USCIS I-9 Central: Form I-9 is required for every person hired for pay; its parts are Section 1 (employee), Section 2 (employer or authorized representative), Supplement A (preparer or translator) and Supplement B (reverification and rehires); Section 2 must be completed within 3 business days of the first day of work for pay (for example, a Monday start means Thursday), or by the first day if the job lasts under 3 days; employees choose one List A document or one List B plus one List C document, and staff must not ask for a specific list; under E-Verify a List B document must have a photo and remote examination is only for E-Verify participants. Output is a timeline per start-date pattern and a checklist of employer duties (examine originals, record document details, enter the physical address and not a P.O. box, return documents).
- HR staff or the named authorized representative examines every document and signs Section 2; the workflow never sees, stores or judges an employee's documents.
- Not legal advice; minors, special placement and unfamiliar documents need USCIS Special Categories and the M-274 handbook, which are flagged rather than summarized. Telling an employee which documents to present is a compliance risk, so the checklist must keep the choice with the employee. Form versions change, so the checklist names the USCIS page review date (Sep 18, 2026).
Sources: Completing Form I-9 | USCIS ↗ (Snapshot pending)Completing Section 2, Employer Review and Attestation | USCIS ↗ (Snapshot pending)Under review conditionalTakes an investigator's draft DMS Plan text (no participant data) and drafts a gap list against NIH's Writing a Data Management and Sharing Plan page: the 2026 DMS Plan format is now required for all competing and non-competing awards; it asks yes or no questions on maximum appropriate sharing, timing (by publication or by the end of the period of performance), and availability for as long as repository or journal policies require; any "no" needs a stated ethical, legal or technical reason in 300 words or fewer; it asks about participant privacy protections and access controls, a table of key data types and repositories (100 words maximum), and Genomic Data Sharing questions where they apply. The gap list also flags reasons NIH says are generally not justifiable (data too small, not widely used, no suitable repository) and reminds staff that a summary and costs go in the budget justification.
Human in the loop: Investigator writes every answer; research administrator reviews the gap list before submission. The workflow never decides what data may be shared. Limits: Not a substitute for the funding Institute or Center's own expectations, which NIH says can add standards or repository rules. Consent and genomic certification questions stay with the IRB and investigator. NIH notes the format was adjusted on Oct 6, 2026, so the check names the page version used. No participant data as inputs. Methodology 1.0: Strong candidate for replay.
How to run: Check a draft NIH Data Management and Sharing Plan against the required 2026 format elements
- Takes an investigator's draft DMS Plan text (no participant data) and drafts a gap list against NIH's Writing a Data Management and Sharing Plan page: the 2026 DMS Plan format is now required for all competing and non-competing awards; it asks yes or no questions on maximum appropriate sharing, timing (by publication or by the end of the period of performance), and availability for as long as repository or journal policies require; any "no" needs a stated ethical, legal or technical reason in 300 words or fewer; it asks about participant privacy protections and access controls, a table of key data types and repositories (100 words maximum), and Genomic Data Sharing questions where they apply. The gap list also flags reasons NIH says are generally not justifiable (data too small, not widely used, no suitable repository) and reminds staff that a summary and costs go in the budget justification.
- Investigator writes every answer; research administrator reviews the gap list before submission. The workflow never decides what data may be shared.
- Not a substitute for the funding Institute or Center's own expectations, which NIH says can add standards or repository rules. Consent and genomic certification questions stay with the IRB and investigator. NIH notes the format was adjusted on Oct 6, 2026, so the check names the page version used. No participant data as inputs.
Sources: Writing a Data Management and Sharing Plan | NIH Grants & Funding ↗ (Snapshot pending)