Add secureagentics/Adrian

[computer-agent]

Adds Adrian by secureagentics to the registry.

What it is: Open-source runtime security monitoring and control engine for AI agents — detects prompt injection, tool-call abuse, policy drift, and misaligned behaviour by analysing both activity logs and reasoning traces. Comes with a Python SDK (two-line LangChain/LangGraph install) and a fully self-hostable Docker stack.

• Repo: https://github.com/secureagentics/Adrian ⭐ 128
• GAP PR (agent.yaml + SOUL.md): Add GitAgent Protocol support (agent.yaml + SOUL.md) secureagentics/Adrian#35
• License: Apache-2.0
• Category: security
• Tags: runtime-security, prompt-injection, agent-monitoring, policy-drift, tool-auditing, langchain, langgraph, mcp, aarm, self-hosted

Note: The registry CI checks for agent.yaml + SOUL.md on the target repo's default branch. Those files are proposed in PR #35 above. Once that PR is merged, all CI checks here will pass automatically.

---

If GAP looks useful, the project lives at ⭐ https://github.com/open-gitagent/opengap — a star helps more maintainers find the standard.