Back to directory
SecurityLGD-Eval/lgd-gate
Lgd Gate
One score, one permission tier. Static security + quality gate for AI agent skills (Claude Skills / MCP). Detects credential exfiltration, base-URL hijack, shell payloads.
Suggested install command
npx skills add LGD-Eval/lgd-gate/Lgd GateAlways inspect the linked repository and skill instructions before running commands. Skills are instructions; permissions and execution still matter.
Compatibility
Agent support matrix
3 supported
| Agent | Status |
|---|---|
| Claude Code | Supported |
| OpenCode | Not listed |
| Cursor | Not listed |
| MCP | Not listed |
| GitHub Copilot | Supported |
| Windsurf |