# panguard-ai

> Open-source security platform for AI agents -- audits skills before install, monitors 24/7, shares threat intelligence across all users. |  AI Agent 開源安全平台 -- 安裝前審計 skill、24/7 即時監控、社群共享威脅情報。

- **URL**: https://www.freshcrate.ai/projects/panguard-ai
- **Author**: panguard-ai
- **Category**: MCP Servers
- **Latest version**: `v1.5.6` (2026-05-10)
- **License**: MIT
- **Source**: https://github.com/panguard-ai/panguard-ai
- **Homepage**: https://panguard.ai
- **Language**: TypeScript
- **GitHub**: 38 stars, 5 forks
- **Registry**: github
- **Tags**: `ai-agent`, `ai-security`, `cybersecurity`, `llm-security`, `mcp`, `open-source`, `prompt-injection`, `sigma-rules`, `typescript`

## Description

Open-source security platform for AI agents -- audits skills before install, monitors 24/7, shares threat intelligence across all users. |  AI Agent 開源安全平台 -- 安裝前審計 skill、24/7 即時監控、社群共享威脅情報。

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `v1.5.6` | 2026-05-10 | High | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v1.5.5...v1.5.6 |
| `v1.4.19` | 2026-04-14 | High | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v1.4.16...v1.4.19 |
| `v1.4.16` | 2026-04-09 | High | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v1.4.13...v1.4.16 |
| `v1.4.13` | 2026-03-27 | Medium | ## What's New  - **Telemetry**: Opt-in anonymous telemetry with first-run consent prompt - **pga up**: Opens dashboard → scans skills → warns about threats → starts protection - **pga status**: Shows all installed skills with security status (SAFE/UNKNOWN) - **pga config set**: Telemetry and Threat Cloud toggle commands - **LLM**: OpenAI-compatible API support (Gemini, Groq, Qwen, DeepSeek via --endpoint) - **TC Metrics**: Real-time whitelist/blacklist/npm download counts - **16 Platform Support |
| `v1.4.2` | 2026-03-27 | Medium | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v1.3.5...v1.4.2 |
| `v1.3.5` | 2026-03-25 | Medium | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v1.2.2...v1.3.5 |
| `v1.2.2` | 2026-03-22 | Low | ## What's New  ### Unified Scanning Engine (`@panguard-ai/scan-core`) - New shared scanning package used by CLI Auditor, Website, and Guard - Unified `patternHash` (`scan:` prefix) enables cross-platform Threat Cloud consensus - CLI now submits ATR proposals for HIGH/CRITICAL findings (flywheel complete) - Website scanner reduced from 885 to 469 lines  ### Bug Fixes - Fixed ESM compatibility for safe-regex import - Fixed contentHash truncation inconsistency between CLI and Website - Fixed risk s |
| `v0.4.5` | 2026-03-15 | Low | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v0.4.4...v0.4.5 |
| `v0.4.4` | 2026-03-15 | Low | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v0.4.3...v0.4.4 |
| `v0.4.3` | 2026-03-15 | Low | **Full Changelog**: https://github.com/panguard-ai/panguard-ai/compare/v0.4.2...v0.4.3 |

## Citation

- HTML: https://www.freshcrate.ai/projects/panguard-ai
- Markdown: https://www.freshcrate.ai/projects/panguard-ai.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/panguard-ai/deps

_Generated by freshcrate.ai. Indexes github releases for AI-agent ecosystem packages._