Guardrail Bridge

• Analyzed: May 12, 2026, 7:19 AM
• Findings: 1
• Version: 1.0.3
• Hash: 8a2da9ec06b2fa59bbfb7bfa606e8fa14f499f1bc03d0dd230f298ec23a640a8
• Source: palm
• External report: View on VirusTotal

Findings (1)

Package: guardrail-bridge (npm) Version: 1.0.3 Description: Pre-agent security plugin for OpenClaw — detects manipulation attempts and blocks policy-violating content before Agent dispatch. The package implements a security guardrail plugin for the OpenClaw platform, designed to check inbound user messages against policy violations using either a local keyword blacklist or remote HTTP moderation services. The code exhibits sound security practices, including: relying on the host platform's SSRF protection mechanisms (`fetchWithSsrFGuard`) for all external HTTP requests; using standard libraries for efficient text processing and keyword matching (`aho-corasick`, Unicode normalization); and implementing robust configuration handling, including specific logic to prevent API key leakage when channel configurations are partially overridden. All observed behavior aligns with the stated purpose of pre-agent content moderation. There are no signs of malicious activities, unauthorized access, or unintended network communication.