This release includes 2 breaking changes for platform teams planning a safe upgrade.
✓ No known CVEs patched in this version
Topics
+14 more
Affected surfaces
Summary
AI summaryCore engine refactor makes ShellWard platform‑agnostic and introduces a new DLP model with audit‑only L2 and block‑only L7 modes.
Full changelog
ShellWard v0.5.0
Major Changes
- Core Engine Refactor: Platform-agnostic engine (
src/core/engine.ts), layers become thin adapters - DLP Model: L2 audit-only (no redaction) — data returns in full for user analysis; L7 blocks external sends
- Bash Exfil Detection: Detects
curl -X POST,wget --post,ncbypass attempts - SDK Mode:
npm install shellward— 3 lines to integrate into any AI agent - Auto Security Check: Scans OpenClaw vulns, plugin risks, MCP config on startup
Supported Platforms
OpenClaw | Claude Code | Cursor | LangChain | AutoGPT | OpenAI Agents | Dify / Coze | Any AI Agent (SDK)
Demo
demo-zh.gif/demo-zh.mp4— 中文版 7 场景演示demo-en.gif/demo-en.mp4— English 7-scenario demo
Tests
112 passing (38 integration + 42 edge cases + 32 SDK)
Breaking Changes
- Core engine refactor: platform‑agnostic design with thin adapter layers replaces the previous monolithic implementation
- DLP model change: L2 now audit‑only (no redaction) and L7 blocks external sends
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About jnMetaCode/shellward
AI Agent Security Middleware & MCP Server with 8-layer defense including prompt injection detection, DLP data flow tracking, command blocking, and PII detection. 7 MCP tools, zero dependencies.
Related context
Beta — feedback welcome: [email protected]