pi-security-scanner
Security scanner and runtime protection for Pi Coding Agent
Package details
Install pi-security-scanner from npm and Pi will load the resources declared by the package manifest.
$ pi install npm:pi-security-scanner- Package
pi-security-scanner- Version
0.1.0- Published
- Feb 27, 2026
- Downloads
- 20/mo · 7/wk
- Author
- burggraf
- License
- unknown
- Types
- extension
- Size
- 44.4 KB
- Dependencies
- 0 dependencies · 0 peers
Pi manifest JSON
{
"extensions": [
"extensions/index.ts"
]
}Security note
Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.
README
Pi Security Scanner
A security extension for pi-coding-agent that provides runtime protection and static analysis for your Pi agent environment.
Features
🛡️ Runtime Shield
- Bash Interceptor: Detects and blocks dangerous bash commands like
curl,wget,nc, and unauthorized system modifications. - File Access Monitor: Protects sensitive files like
.env,.ssh/keys, and.git/configfrom unauthorized writes or edits. /security-shieldCommand: Enables or disables the Runtime Shield. When disabled, no bash commands or file access are intercepted.
🔍 Static Scanner
/security-scanCommand: Scans all installed Pi extensions (globally and locally) for dangerous patterns such aseval(),child_process.exec(), and unauthorized network calls.
Security Check Documentation
For detailed information about all security checks, see docs/security-checks.md. This document explains:
- What patterns are flagged and why
- Severity levels (HIGH, MEDIUM, LOW)
- How to respond to different types of warnings
- What patterns are intentionally NOT flagged (to reduce false positives)
Installation
This extension is configured as a pi-package. You can install it by adding it to your Pi configuration:
pi install npm:pi-security-scanner
How it Works
The scanner leverages Pi's built-in Extension API:
tool_callEvent Hooks: Intercepts tool execution to provide real-time guardrails.- Heuristic Engine: Uses regex-based analysis to identify suspicious code patterns in extension source files.
- User Confirmation: Never blocks silently—always asks for user permission before stopping a suspicious action.