Cogni
AI study app that decides what to study — persistent tutor memory, grade tracking with a 'what do I need on the final' calculator, and study through your own Claude via MCP. BYOK on Next.js + Supabase.
Caution. Specific findings reduced this grade. They are listed on the page. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.
How we verified
Verified is a snapshot. Live keeps it current, and builds your track record.
⚡ Connect GitHub → continuous verification on every pushwhy connect →Who stands behind it
arshawnarbabi
Source: github_repo_search
Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →
Install from
The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.
These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.
list_coursesList the student's active courses (course_id + name). Call this first — the other tools need a course_id.
get_course_overviewTopics (with mastery %), upcoming exams, and pending assignments for a course. Use it to focus tutoring on what the student actually needs.
get_weak_topicsThe student's highest-leverage weak topics (high professor weight × low mastery). Steer practice here.
search_materialsSemantic search over the student's UPLOADED course materials. Call this before explaining a topic and treat the results as authoritative. Returns the most relevant excerpts.
get_due_cardsFlashcards the student has due for review today (optionally for one course). Useful for running a quick review.
create_flashcardsCreate spaced-repetition flashcards for a topic. They are saved to the student's Cogni account and enter their review queue immediately.
get_learning_profileThe durable notes Cogni keeps about how this student learns (strengths, recurring misconceptions). Use it to calibrate your explanations.
review_cardRecord the student's rating for a flashcard (1=Again, 2=Hard, 3=Good, 4=Easy). Updates FSRS scheduling AND topic mastery in Cogni — call this after quizzing the student on each card from get_due_cards, so the review session actually counts (reschedules the card, moves their mastery, keeps the in-app…
grade_answerRecord how well the student understood a topic (0.0–1.0) after they answered a question you posed. Moves their Cogni mastery score — without this, an hour of tutoring changes nothing in their study plan. Use topic_id from get_course_overview / get_weak_topics.
log_study_sessionCall at the END of a tutoring/study session: records it in the student's Cogni history, feeds their cross-session memory, and keeps their study streak alive. Without this, study done here is invisible in the app.
get_study_planThe student's Cogni study plan for today (or a given date): review blocks, homework due, quizzes — with urgency reasons. Use it to answer 'what should I work on?'
complete_assignmentMark a homework assignment done in Cogni (assignment_id from get_course_overview or get_study_plan). Recomputes today's study plan so it reflects the change.
record_quiz_resultAfter YOU quiz the student conversationally (you author the questions — costs them nothing), record the result so it shows in their Progress tab and moves mastery. Provide per-topic scores when you can.
researchMulti-query search over the student's uploaded materials for synthesizing a study guide or deep answer. YOU write 2-5 angled sub-queries (definitions, examples, edge cases, professor's framing); the tool returns deduped excerpts with source filenames for citation.
Disclosed vulnerabilities in this server's declared npm dependencies (via OSV). Whether each is reachable depends on the installed versions.
Next.js: Unauthenticated Remote Code Execution in Image Optimization API when AVIF files are used
Next.js: Unauthenticated Remote Code Execution on windows-hosted servers
Next.js: Middleware / Proxy bypass in App Router applications using Turbopack and single locale
Next.js: Server-Side Request Forgery in Server Actions on custom servers
Next.js: Denial of Service in App Router using Server Actions
ANTHROPIC_KEYAPP_URLCRON_HEARTBEAT_MAINTENANCE_URLHeartbeat alert URL for the maintenance cronCRON_HEARTBEAT_NUDGE_URLHeartbeat alert URL for the nudge cronCRON_HEARTBEAT_SCHEDULER_URLHeartbeat alert URL for the scheduler cronCRON_SECRETLong random secret used by Vercel cron endpointsGOOGLE_CALENDAR_CLIENT_IDGoogle Cloud Console — Calendar OAuthGOOGLE_CALENDAR_CLIENT_SECRETGoogle Cloud Console — Calendar OAuthKV_URLOPENAI_KEYOPERATOR_SECRETBearer secret guarding the operator console + audit endpointsSUPABASE_DB_URLTURNSTILE_SECRET_KEYCloudflare Turnstile CAPTCHA — secret keyNEXT_PUBLIC_PASSWORD_RESET_ENABLEDNEXT_PUBLIC_SENTRY_DSNNEXT_PUBLIC_SUPABASE_ANON_KEYNEXT_PUBLIC_SUPABASE_URLNEXT_PUBLIC_TURNSTILE_SITE_KEYREDIS_URLSENTRY_AUTH_TOKENSENTRY_DSNSENTRY_ORGSENTRY_PROJECTSUPABASE_SERVICE_ROLE_KEYSUPABASE_URLTool annotations
No tools have read-only/destructive annotations
Add readOnlyHint or destructiveHint annotations to every tool so hosts can warn users before invoking.
All four hints declared on every tool
14/14 tools missing one or more hints — list_courses (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_course_overview (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_weak_topics (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint), +11 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.
For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.
Tool test coverage
Only 1/14 tools referenced in tests (7%)
Write tests that reference each tool by name so every tool has at least one test.
Secrets stay with their owner
1 secret/sensitive value flow into network calls (TURNSTILE_SECRET_KEY → challenges.cloudflare.com) (1 other flows matched canonical API hosts)
Audit where credentials are sent. A NOTION_TOKEN should only reach api.notion.com — never a third-party host.
Production dependencies are patched
2 critical, 4 high severity in production deps — next@16.2.9 (critical), next@16.2.9 (critical)
Run npm audit fix, or upgrade the affected packages to a non-vulnerable version.
Dependency freshness
2/50 production deps stale: remark-math@2023-11-20 (2.8y), rehype-katex@2024-08-19 (2.1y)
Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.
[](https://m8ven.ai/mcp/arshawnarbabi/cogni)?variant=verified from the URL.Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.
https://m8ven.ai/api/mcp/tool-check