Brisk Teaching MCP server icon

Brisk Teaching

by Brisk Teaching

HIPAA CompliantSOC2 ReadyISO 27001 Ready
Data & Research9 tools

Create Boost activities, lesson plans, quizzes and standards-aligned materials in a Brisk account from an AI agent. Nine tools, all authoring. Brisk states one tool reads class names but never student data. OAuth sign-in, identity-only scopes.

Verified connector

Listed by Anthropic as a partner connector in its Connectors Directory.

Connection checked by Agentman on .

Anthropic states this reflects the level of review a connector received, not a security audit.

Connect Brisk Teaching via MCP

https://mcp.briskteaching.com/mcp

Works in any MCP-compatible client. In Agentman Studio it is one click — no config file to edit.

Brisk Teaching Tools & Capabilities (9)

Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.

Limits

  • No application scopes exist. Brisk's authorization server advertises four scopes — email, offline_access, openid and profile — every one an OpenID identity claim, and the RFC 9728 resource descriptor declares scopes_supported as an empty array. The OAuth grant offers no way to permit reading standards without also permitting authoring in the teacher's account.
  • The student-data boundary is documented, not schema-verified. The endpoint returned 401, so no parameter schemas were readable. Brisk's "never student data" statement and the batch-feedback split are the vendor's prose. Every source we checked agrees with them and none contradicts them, but we did not confirm them from the wire.
  • Three of nine tool names matched only semantically. finalize_boost_activity, generate_next_ideas and create_batch_feedback_assignment map to Brisk's labels by meaning rather than by string, so a rename within those three would be invisible to this diff.
  • Some tools spend credits and nothing can observe the balance. Brisk states batch feedback and quizzes draw on free-plan credits. No tool reads a credit, quota or usage figure, so an agent cannot check the remaining balance before spending it or report what a run cost.
  • Batch Feedback is beta. Brisk's guide opens by stating the feature "is still in Beta", that it may change frequently, and that users "may encounter a bug".
  • Google is a hard dependency for document output. Without a Google account connected in Brisk, the tools that emit a Google Doc or Google Form cannot deliver their result.
  • Writes land in a real teacher account. Every create tool acts inside the signed-in Brisk account, and no tool deletes or archives. An over-eager agent produces clutter a teacher must clear by hand in the Brisk web app.
  • Whether an activity link is guessable is unresolved. Brisk states students join a Boost activity with a class code or link and no account, signing in with Google or simply adding a name. We found no documentation of how those links or codes are scoped, and we did not test it.

Frequently asked questions

No, and Brisk states the limit explicitly. Its connector documentation describes get_teacher_context as reading your class names, grades and subjects, and adds the parenthetical never student data. Grades there means grade levels such as 7th grade, not marks. No tool among the nine returns a submission, a score or a named student.

No. create_batch_feedback_assignment creates only an assignment shell. Brisk's documentation states you add your students' work in the Brisk web app, then run the AI feedback there. Importing submissions requires Google Drive or Google Classroom inside the extension, and Brisk requires teachers to view feedback directly before approving it.

Not to students. Brisk states activities start as drafts and that creating a Boost activity never publishes it to students. The connector's finalize_boost_activity marks setup complete so an activity is ready to share. Distribution is a separate step in the Brisk web app, where the teacher shares a link, QR code or class code.

Brisk states it complies with FERPA, COPPA, GDPR and Canadian provincial privacy legislation, and describes its Boost technology as SOC 2, FERPA and COPPA compliant. Brisk also states it does not use student or teacher data to train or fine-tune AI models and does not sell or share personal data. Those are Brisk's claims about the platform.

Some tools do. Brisk's connector documentation states a few tools, naming batch feedback and quizzes, use free credits on the free plan, and that paid plans remove these credit limits. No tool among the nine can read a credit balance or quota, so an agent cannot check remaining credits before spending them.

Only identity claims. Brisk's authorization server advertised four scopes on 2026-08-23: email, offline_access, openid and profile. The server's own RFC 9728 descriptor declares scopes_supported as an empty array. So nothing at the OAuth grant separates creating a lesson plan from creating a batch feedback assignment.

A Claude account, a Brisk Teaching account, and a connected Google account. Brisk states you sign in with the same account you use at app.briskteaching.com, and that any tool producing a Google Doc or Google Form, such as lesson plans, worksheets and quizzes, needs a Google account connected in Brisk first.

No, it looks them up. Brisk states that Claude finds matching standards in Brisk's database and confirms them with you before applying them, so materials reference real official standard codes. The list_standards tool browses that database by framework, state and grade rather than generating codes from the model.

Sources

Use in Agentman

Connect once and your agents call these tools on their own — on a schedule, in a workflow, with nobody at the keyboard.

Open in Agentman Studio

Server Info

Category
Data & Research
Developer
Brisk Teaching
Tools
9
Domain
mcp.briskteaching.com

Using Claude Desktop or another MCP client? Setup docs — the connection URL above works anywhere.