Sentris: A Security Scanner for Supabase & AI-Written Code

Research context and background

Sentris is a security scanner built specifically for applications that use Supabase and code written by artificial intelligence. It helps developers find dangerous mistakes in their code before hackers can exploit them. The tool reads both the code files in a developer's repository and the live website bundle to spot issues like leaked secret keys, missing security rules, and open storage buckets. Unlike older tools that only look at a website URL, Sentris dives deep into the actual code structure to provide precise fixes.

Benefits

Sentris offers several key advantages for developers who want to keep their applications safe. First, the initial scan is completely free and requires no login. Users can get results in just a few seconds without signing up for an account. The tool is read-only, meaning it never clones or stores files permanently. It only reads the content during the scan and discards it afterward. Every finding includes the exact file name and line number where the problem exists. It also provides copy-paste SQL or code fixes to resolve the issue quickly. The tool uses a strict verification process to ensure zero false positives, so developers do not waste time on incorrect alerts. AI findings are clearly labeled as potential issues so they do not trigger unnecessary monitoring alarms.

Use Cases

Sentris is designed for developers building modern web applications with Supabase. It is especially useful for teams that rely heavily on AI coding agents to write their code. Developers can integrate Sentris with tools like Cursor or Windsurf to run scans directly from their code editor. The tool helps identify three common mistakes that often lead to data breaches. These include leaked service role keys that bypass security rules, tables without Row Level Security policies, and public storage buckets that anyone can access. It also checks for missing security headers, weak password hashing, and unvalidated user input. The tool is suitable for individual side projects as well as larger teams that need scheduled monitoring and history tracking for multiple repositories.

Pricing

Scanning itself is free for everyone. However, advanced features like the AI verification pass and scheduled monitoring require a subscription. The Starter plan costs $19 per month and allows one connected target with unlimited re-scans and weekly monitoring. The Pro plan is $49 per month and supports up to five targets with nightly monitoring and scan history. The Team plan costs $99 per month and offers unlimited targets across entire organizations with nightly monitoring.

Vibes

Sentris has received positive attention for its transparency and accuracy. The tool explicitly states which files it read and which it skipped during a scan. This coverage ledger builds trust with users who want to know exactly how the analysis was performed. The fact that it reported zero false positives in tests with planted vulnerabilities demonstrates its reliability. Developers appreciate that it does not require a sales call to get started, making it accessible for small teams and solo creators. The clear separation between confirmed issues and potential AI findings helps reduce noise in the developer workflow.

Additional Information

Sentris operates with a focus on privacy and security. It never uses a developer's repository for training machine learning models. File contents are sent to a model provider only when the AI verification pass is enabled, and this feature can be turned off to use only deterministic rule checks. The tool integrates with AI coding agents via an MCP server to automate the scanning process. It distinguishes between confirmed security issues based on strict rules and potential issues verified by an AI pass. This approach ensures that the tool remains accurate while leveraging the power of artificial intelligence to catch complex problems.