Cursor Privacy SettingsSAFE
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Overview
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
4f83675ca38aOBSERVED · 2026-10-08Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned | |
| cursor | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: cursor-privacy-settings description: 'Configure Cursor privacy mode, data handling, telemetry, and sensitive file exclusion. Triggers on "cursor privacy", "cursor data", "cursor security", "privacy mode", "cursor telemetry", "cursor data retention". ' allowed-tools: Read, Write, Edit, Bash(cmd:*) version: 1.19.0 license: MIT author: Jeremy Longshore <[email protected]> tags: - saas - cursor - security compatibility: Designed for Claude Code --- # Cursor Privacy Settings ## Overview Configure Cursor data, privacy, and model-routing settings according to the repository's data classification and organization policy before code is added to AI context. ## Prerequisites - A current data-classification decision, tenant owner, and approved provider/model policy. - Authority to inspect or enforce tenant settings and a process for exceptions/incidents. ## Instructions 1. Select the approved privacy/data controls and verify their effective tenant/user state. 2. Configure `.cursorignore` and workspace trust to prevent secrets, customer data, and generated artifacts entering context. 3. Test the boundary with sanitized fixtures and retain redacted verification evidence. 4. Review settings after tenant, provider, policy, or data-classification changes. ## Output - A verified privacy configuration and context-exclusion boundary for the intended workspace. ## Error Handling | Condition | Safe response | |---|---| | Effective privacy setting is unclear | Do not use sensitive code; escalate to the tenant/security owner. | | Sensitive path is included | Stop, remove/exclude it, and follow the exposure procedure. | | Required setting conflicts with policy | Use the restrictive option and obtain a documented exception decision. | ## Examples Before opening a production repository, verify the approved privacy mode, add known secret/export paths to reviewed ignore configuration, and test with a sanitized path. If the file can still be attached, stop onboarding the repository until the boundary is fixed. Configure Cursor's privacy controls to protect your code and data. Covers Privacy Mode, data handling policies, file exclusion, telemetry, and enterprise security settings. ## Privacy Mode ### What Privacy Mode Does | With Privacy Mode ON | With Privacy Mode OFF | |---------------------|----------------------| | Zero data retention at model providers | Providers may retain data per their policies | | Code not used for training (Cursor or providers) | Code may be used to improve AI models | | Embeddings computed without storing source | Same embedding behavior | | Telemetry: anonymous usage only | Telemetry may include code snippets | ### Enabling Privacy Mode **Individual:** `Cursor Settings` > `General` > `Privacy Mode` > ON **Team enforcement (Business/Enterprise):** Admin Dashboard > Privacy > "Enforce Privacy Mode for all members" When team-enforced: - Individual users cannot disable Privacy Mode - Client pings server every 5 minutes to verify enforcement - New members automatically have Privacy Mode enabled ### Verifying Privacy Mode 1. `Cursor Settings` > `General` -- check Privacy Mode toggle 2. [cursor.com/settings](https://cursor.com/settings) -- shows account-level status 3. For teams: Admin Dashboard shows enforcement status per member ## Data Flow: Where Your Code Goes ``` Your Code in Editor │ ├─► Tab Completion ──► Cursor's proprietary model server │ (zero retention with Privacy Mode) │ ├─► Chat/Composer ──► Model provider (OpenAI/Anthropic/Google) │ (zero retention agreements in place) │ ├─► Codebase Index ─► Cursor embedding API ─► Turbopuffer (vector DB) │ (embeddings only, no plaintext code) │ └─► BYOK ───────────► Your API provider directly (your provider's data policy applies) ``` ### What IS Stored | Data | Stored Where | Retention | |------|-------------|-----------| | Embeddings (vectors) | Turbopuffer (cloud) | Until project re-indexed | | Obfuscated file metadata | Cursor servers | Active session only | | Anonymous telemetry | Cursor analytics | Aggregated, no PII | | Account info | Cursor auth servers | While account active | ### What Is NOT Stored (Privacy Mode ON) - Plaintext source code - Chat prompts and responses - File contents sent for completion - Code snippets from Tab suggestions ## Sensitive File Exclusion ### .cursorignore (Best-Effort AI Exclusion) ```gitignore # .cursorignore -- prevent files from AI features + indexing # Credentials and secrets .env .env.* .env.local .env.production **/secrets/ **/credentials/ **/*.pem **/*.key **/*.p12 # Regulated data **/pii/ **/hipaa/ **/financial-data/ # Internal configuration .cursor-config-private infrastructure/terraform.tfvars ``` **Important:** `.cursorignore` is best-effort. Due to LLM unpredictability, it is not a hard security boundary. Do not rely solely on `.cursorignore` to protect truly sensitive data. ### Defense in Depth ``` Layer 1: .gitignore → Secrets never in repo Layer 2: .env files → Config via environment variables Layer 3: .cursorignore → Best-effort AI exclusion Layer 4: Privacy Mode → Zero data retention at providers Layer 5: BYOK + Azure → Route through your own infrastructure ``` ## Telemetry Configuration ### What Cursor Collects With Privacy Mode ON, telemetry is limited to: - Feature usage counts (how often Chat/Composer/Tab used) - Error reports (crashes, not code content) - Performance metrics (response times) - Extension compatibility data ### Disabling Telemetry ```json // settings.json { "telemetry.telemetryLevel": "off" } ``` Or: `Cursor Settings` > search "telemetry" > set to "off" **Note:** Disabling telemetry may reduce Cursor's ability to diagnose issues affecting your account. ## Network Security ### Required Domains Allowli
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
4f83675ca38afull audit observations/trust-audit/skill/jeremylongshore__cursor-privacy-settings.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 4f83675ca38a | SAFE | B | 89 | first audit |
Questions
What does the Cursor Privacy Settings skill do?
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Is Cursor Privacy Settings safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Cursor Privacy Settings access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does Cursor Privacy Settings work with?
Its documentation mentions claude-code and cursor. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (4f83675ca38a), read on 2026-10-08. The repository is watched, and a new audit runs when it changes — this is the first audit.