By websentry-ai
Use Cursor without fear of data leakage. Unbound enforces command policies, DLP and jailbreak guardrails, and streams session analytics to your security dashboard.
Security, compliance, and observability for Cursor — powered by Unbound.
Unbound monitors every AI-driven action in Cursor in real time — shell commands, MCP tool calls, file reads and edits, user prompts, and assistant responses. It enforces the policies and guardrails you configure in the Unbound dashboard, and streams full session data for analytics and audit.
Key capabilities:
From Cursor Marketplace
Search for Unbound in Settings > Plugins, or visit cursor.com/marketplace.
From source
git clone https://github.com/websentry-ai/cursor-extension.git
cd cursor-extension
./install.sh
The installer opens your browser automatically. Sign in to your Unbound account and the API key is saved to your shell profile.
Or run the setup skill inside Cursor:
/unbound-setup
After setup, try these in Cursor:
ls — check your Unbound dashboard for the event| Event | When it fires | What Unbound does |
|---|---|---|
| Shell command | Before execution | Checks against your command policies |
| MCP tool call | Before execution | Checks against your tool policies |
| User prompt | Before submission | Runs DLP, NSFW, and jailbreak guardrails |
| File read | Before read | Logs to audit trail |
| File edit | After edit | Logs to audit trail |
| Shell output | After execution | Logs to audit trail |
| MCP result | After execution | Logs to audit trail |
| Agent response | After response | Captures for session analytics |
| Session end | On stop | Sends full exchange to dashboard |
All checks happen in real time. Blocked actions show an explanation to the user.
The only required configuration is your API key:
| Variable | Description |
|---|---|
UNBOUND_CURSOR_API_KEY | Your Unbound API key. Get one at app.getunbound.ai > Settings > API Keys |
The installer and /unbound-setup skill handle this automatically. The key is saved to your shell profile (~/.zprofile, ~/.bashrc, etc.) and picked up by Cursor on launch.
For fleet-wide enforcement where users cannot disable the plugin:
See enterprise/README.md for detailed instructions.
./uninstall.sh
Or remove the plugin from Cursor's Settings > Plugins panel.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub websentry-ai/cursor-pluginUse Claude Code without fear of data leakage. Unbound enforces command policies, DLP and jailbreak guardrails, and streams session analytics to your security dashboard — all with zero-latency fail-open design.
A secure runtime for Claude Code. Intercepts every tool call with policy-based allow/block/ask decisions, evasion detection, path fencing, file snapshots, and audit logging.
Achieve flow state safely with Claude Code. Auto-approves routine work, gates risky actions, hard-blocks dangerous patterns. Dual enforcement (skill + hooks), token cap for cost governance, full audit trail. Zero dependencies.
Smart command safety filter for Claude Code — parses shell pipelines and evaluates per-command safety rules to auto-approve safe commands and block dangerous ones
Command Line Agent Safety Harness. All interactions with clash policy should go through this plugin
Pre-execution safety layer that blocks dangerous shell commands and credential file reads using pattern matching + LLM analysis. Fail-closed design.
Protect secrets and credentials from Claude Code. Blocks writes to .env files, detects API keys in shell commands, prevents hardcoded tokens, and guards service account JSON files.