Your AI coding agent writes the code. VibeGuard checks the security.
VibeGuard AI is an open-source security scanner designed for modern applications built with AI coding tools, no-code/low-code platforms, and vibe-coding workflows.
It scans your project source code for potential security issues and provides clear findings, confidence levels, and recommendations.
- π Security vulnerability scanning
- π¨ Severity-based findings
- π― Confidence score for findings
- π‘οΈ Potential XSS detection
- π Authentication & authorization checks
- π Secret and credential detection
- π API security checks
- π¦ Dependency security analysis
- βοΈ Security configuration checks
- π€ AI-assisted security analysis
- π Multiple report formats (JSON, Markdown, HTML, SARIF)
- π Git diff scanning (
--diff) - π» CLI-based workflow
- π VS Code / Cursor integration
- π§βπ» Designed for vibe-coded applications
Before installing VibeGuard, make sure you have:
- Git
- Node.js
- npm
Check your installation:
git --version
node --version
npm --versionClone VibeGuard from GitHub:
git clone https://github.com/speakcoreofficial-stack/VibeGuard-AI-Pluging.gitEnter the project:
cd VibeGuard-AI-PlugingCheck the project files:
dirOn macOS/Linux:
lsInside the VibeGuard-AI-Pluging directory, run:
npm installWait until the installation completes successfully.
Build the CLI:
npm run buildAfter the build completes, verify that the CLI exists.
Windows PowerShell:
Test-Path .\packages\cli\dist\index.jsThe expected result is:
True
If the result is True, the VibeGuard CLI has been successfully built.
You can test VibeGuard directly without installing the global command.
From the VibeGuard repository:
node ./packages/cli/dist/index.js versionIf a version number is displayed, the CLI is working correctly.
Example:
VibeGuard 0.1.0
VibeGuard is designed to scan your own project or a project you are authorized to test.
For example, suppose your project is located at:
D:\aditya\my-project
Move into your project:
cd D:\aditya\my-projectThen run VibeGuard:
node D:\aditya\VibeGuard-AI-Pluging\packages\cli\dist\index.js scanVibeGuard will analyze the project and report potential security findings.
VibeGuard may report findings like:
[HIGH] Potential XSS: Unsafe HTML rendering
File: src/components/Example.tsx
Confidence: 70%
Recommendation:
Use safe rendering methods or sanitize untrusted HTML
before rendering.
Each finding can include:
- Severity
- Vulnerability type
- File location
- Confidence score
- Recommendation
- Evidence
vibeguard scanvibeguard scan --file src/auth.tsvibeguard scan --diffvibeguard scan --aivibeguard scan --civibeguard scan --format json
vibeguard scan --format markdown
vibeguard scan --format html
vibeguard scan --format sarif
vibeguard scan --format json -o report.jsonvibeguard initNote: The
vibeguardcommand requires the CLI to be installed globally. If it is not available on your system, use the direct Node.js command shown in the installation section.
If you want to use:
vibeguard scaninstead of:
node ./packages/cli/dist/index.js scaninstall VibeGuard globally from the cloned repository (run from the packages/cli directory):
cd packages/cli
npm install -g .Then verify:
vibeguard versionExpected output:
VibeGuard 0.1.0
If the command is recognized, you can scan your project with:
cd path/to/your-project
vibeguard scanReplace path/to/your-project with the actual path of your project.
VibeGuard is designed to work with VS Code and Cursor.
Build the extension:
cd packages/vscode
npm installInstall the VS Code extension packaging tool:
npm install -g @vscode/vsceCreate the VSIX package:
vsce packageThis will generate a .vsix file.
In VS Code or Cursor:
Extensions
β
...
β
Install from VSIX
β
Select the generated .vsix file
After installation, VibeGuard can be used directly from the editor.
VibeGuard includes plugins so your AI coding agent scans automatically.
npm run build -w @vibeguard/integrations
cp packages/integrations/dist/opencode/opencode.js .opencode/plugin/opencode.jsOr reference it from opencode.json:
{
"$schema": "https://opencode.ai/config.json",
"plugin": ["/absolute/path/VibeGuard-AI-Pluging/packages/integrations/dist/opencode/opencode.js"]
}What you get:
- A
vibeguard_scantool the agent can call - A
/vibeguard-scancommand for a readable summary + fix discussion - An auto security hint after every
edit/writethat introduces findings
Restart opencode after adding the plugin.
mkdir -p .claude-plugin
cp -r packages/integrations/claude/. .claude-plugin/What you get:
/vibeguardcommand β runsnpx vibeguard scanand explains findingsvibeguard-scanskill β automatically triggers when generating code or when asked "is this code secure?"post_tool_use.shhook β post-edit security hint for source files
VibeGuard includes example applications for testing the scanner.
Navigate to the example:
cd examples/vulnerable-appThen run:
node ../../packages/cli/dist/index.js scanThis allows developers to test VibeGuard locally without scanning an external application.
Expected result (intentionally vulnerable example):
Critical: 1 High: 4
Security Score: 40/100
[CRITICAL] Possible exposed API credential
[HIGH] Vulnerable dependency detected: express
[HIGH] Vulnerable dependency detected: axios
[HIGH] Potential XSS: Unsafe HTML rendering
[HIGH] Potential Missing Authorization
VibeGuard-AI-Pluging/
β
βββ packages/
β βββ core/ # @vibeguard/core β scanners, findings, reporter
β βββ cli/ # vibeguard β the CLI
β βββ vscode/ # VS Code / Cursor extension
β βββ ai/ # @vibeguard/ai β AI reviewer
β βββ integrations/ # opencode + Claude Code plugins
β
βββ examples/
β βββ vulnerable-app/
β
βββ README.md
βββ LICENSE
βββ package.json
VibeGuard is designed as a defensive security tool.
It focuses on helping developers identify and fix security problems in applications they own or are authorized to test.
VibeGuard should not be used to access, attack, or test systems without permission.
VibeGuard is local-first: your source code is scanned on your machine by default and never uploaded unless you explicitly enable an AI provider.
Modern AI coding tools can generate applications extremely quickly.
However, AI-generated code can also introduce:
- insecure authentication
- missing authorization
- exposed secrets
- unsafe input handling
- insecure APIs
- vulnerable dependencies
- unsafe configuration
- XSS risks
VibeGuard acts as a security layer between AI-generated code and production.
AI Coding Agent
β
Generated Code
β
π‘οΈ VibeGuard
β
Security Analysis
β
Findings + Recommendations
β
Safer Application
- CLI foundation
- Source-code scanning
- Security findings
- Severity levels
- Confidence scores
- Basic XSS detection
- Secret detection
- Dependency scanning
- Single-file scanning (
--file) - Git diff scanning (
--diff) - AI explanations (
--ai) - CI mode (
--ci) - JSON / Markdown / HTML / SARIF reports
- VS Code + Cursor extension
- opencode + Claude Code plugins
- Improved false-positive filtering
- Finding deduplication
- Authentication analysis
- GitHub Actions
- Pull Request security comments
- Python support
- Go support
- Java support
- PHP support
- Advanced data-flow analysis
- AI security agent
- Architecture analysis
- Runtime verification
- Automated security verification
- Continuous security monitoring
Contributions are welcome!
git clone https://github.com/speakcoreofficial-stack/VibeGuard-AI-Pluging.gitgit checkout -b feature/my-feature4. Test your changes
Scan report Β· 2026-09-17
- β Prohibited terms or links
- β Repository eligibility
- β slopscore.md paperwork
- β Content policy
- β Risk review
From the balcony Β· 0 of 4 clapped
Princess, Crusoe, Schnitzel and Cap'm Slop read it and passed. Their reasons are on the balcony, with every other verdict.
Critics are accounts on this site with no GitHub account behind them. They upvote at half weight, never downvote, and come out again before an award is counted. Who they are.
0 comments
log in to comment.