Student Records MCP Server

A full Student Information System (SIS) for ADK-Rust Enterprise education agents. 46 MCP tools covering students, courses & enrollment, grades & rubric scoring, attendance, support cases, guardians & communications, transcripts, and learning analytics — with FERPA access logging on every student-scoped operation.

A platform, not a point solution

This is modeled as a general SIS (à la Banner / PowerSchool / Workday Student), so a wide range of education agents are simply clients: Academic Policy Advisor, Adaptive Tutor, Learning Analytics, Parent Communication, Rubric Grading Assistant, and Student Support — plus anything else that needs to read or update the student record under proper governance.

Architecture

Student Records MCP Architecture

Capabilities

Students — demographics, enrollment status, accommodations (IEP/504/ELL)
Courses & sections — catalog + term offerings with grade-category weights
Enrollment — student↔section with status
Grades & rubrics — assignment grades, rubric scoring (criterion sums), and a weighted gradebook → letter grade
Attendance — records + computed rate
Support — intervention cases with notes and status
Guardians & communications — guardian links with consent; messaging
Transcript — completed-course history + cumulative GPA (4.0 scale)
Learning analytics — risk standing derived from grade average + attendance

FERPA posture

Every student-scoped read or write is logged to a per-student access log (accounting of disclosures); get_access_log returns it.
Attributable access — every student-scoped tool requires an actor.
Guardian communications are consent-gated — send_communication refuses unless the guardian belongs to the student and has consent on file; it's classed external_write and gated.
Minimal PII — guardians carry a contact reference, not raw contact details; sample data is fictitious.

Tools (27)

Students (5)

find_students · get_student · set_student_status (gated) · add_accommodation (gated) · get_access_log

Courses, Sections & Enrollment (5)

create_course · list_courses · create_section · enroll (gated) · get_enrollments

Grades & Gradebook (4)

create_assignment · record_grade (gated; rubric-aware) · get_grades · gradebook

Attendance (2)

record_attendance (gated) · attendance_summary

Support (5)

open_support_case · get_support_case · add_case_note · set_case_status · get_support_cases

Guardians & Communications (4)

add_guardian (gated) · get_guardians · send_communication (FERPA-gated, external) · get_communications

Records & Analytics (2)

get_transcript · analytics

Competency & Mastery — Adaptive Tutor (5)

add_competency · list_competencies · record_mastery_evidence (gated) · get_mastery · learning_path

Mastery is tracked per learning standard (with prerequisites). learning_path recommends the competencies a student is ready to learn next — not yet proficient, prerequisites met — and what's blocked, giving an Adaptive Tutor real mastery data instead of letter grades.

Degree Audit & Standing — Academic Policy Advisor (7)

add_grad_requirement · list_grad_requirements · degree_audit · academic_standing · place_hold (gated) · release_hold (gated) · get_holds

degree_audit reports earned-vs-required credits per subject for the student's program (progress %, on-track flag); academic_standing derives honor-roll/probation from GPA; holds gate registration/transcript/financial actions.

Early Warning & Interventions — Student Support / Learning Analytics (7)

raise_flag · set_flag_status · get_flags · evaluate_early_warning · assign_intervention (gated) · end_intervention · get_interventions

evaluate_early_warning auto-raises flags from grades, attendance, and mastery; interventions follow MTSS/RTI tiers (1–3).

Example

> find_students(actor: "ms.carter", name: "mwangi")        → STU-1001 Mwangi, Aisha
> gradebook(actor: "ms.carter", student_id: "STU-1001", section_id: "SEC-…")
  → overall 84.4% = B   (HW 90% × 0.3 + Midterm 82% × 0.7)

> analytics(actor: "counselor", student_id: "STU-1001")
  → grade_average 86 · attendance_rate 100 · risk_level "low"

> send_communication(student_id: "STU-1001", guardian_id: "GRD-…", subject: "Progress", body: "…", sent_by: "parent.agent")
  → COM-1010   (allowed: guardian has consent on file)

> get_access_log(actor: "audit", student_id: "STU-1001")
  → 13 entries — full accounting of disclosures

Rubric grading:

> record_grade(assignment_id: "ASG-…", student_id: "STU-1001",
    rubric: [{name:"Thesis",max_points:10,awarded:8}, {name:"Evidence",max_points:10,awarded:9}])
  → points_earned 17 (sum of awarded criteria)

Installation

1. Build

git clone https://github.com/zavora-ai/mcp-student-records
cd mcp-student-records
cargo build --release

2. Add to your MCP client

Claude Desktop / Kiro / Cursor / Windsurf:

{
  "mcpServers": {
    "student-records": {
      "command": "/path/to/mcp-student-records"
    }
  }
}

3. Use it

> find_students(actor: "advisor", grade_level: "10")
> get_transcript(actor: "registrar", student_id: "STU-1001")

Data Handling

FERPA by construction — access logging, attributable access, and consent-gated guardian communications are built in.
This is an integration scaffold, not a certified SIS. In production, back it with your real SIS of record and enforce auth, RBAC (role-appropriate fields), and encryption; bind actor to an authenticated identity.
No real student PII ships in the crate; sample data is fictitious.

MCP Server Manifest

server_id = "mcp_student_records"
display_name = "Student Records (SIS)"
version = "1.1.0"
domain = "education"
risk_level = "high"
writes_allowed = "gated"
transports = ["stdio"]

Contributors

|
_{James Karanja Maina} | |:---:|

License

Apache-2.0 — see LICENSE for details.

Part of the ADK-Rust Enterprise MCP server ecosystem.

Registry Compliance

This server implements the ADK MCP SDK contract:

HealthCheck — async health probe for registry monitoring
mcp-server.toml — manifest declaring tools, risk classes, and approval gates
Structured tracing — RUST_LOG env-filter for observability

MCP Servers