feat: add AgentSight and ActPlane (eBPF observability/security)

- AgentSight (Agent Evaluation & Observability): zero-instrumentation
  eBPF observability for LLM/coding agents
- ActPlane (Agent Security): OS-level agent harness compiling policy DSL
  to eBPF for syscall-boundary information-flow control

Co-authored-by: LinuxDev9002 <linuxdev8883@example.com>

Author: yunwei37

README.md

@@ -586,6 +586,7 @@ Entries may carry one or more status tags so readers can judge maturity at a gla
 - [RAMPART](https://github.com/microsoft/RAMPART) - 🆕 **May 20, 2026**. Microsoft's pytest-native safety + security testing framework for agentic AI. Developer-facing white-box counterpart to PyRIT — cross-prompt-injection probes, benign-failure asserts, harm-category coverage, statistical thresholds (e.g. safe in 80%+ runs). Integrates straight into CI/CD. MIT. ![GitHub stars](https://img.shields.io/github/stars/microsoft/RAMPART?style=flat-square)
 - [Clarity (Microsoft)](https://www.microsoft.com/en-us/security/blog/2026/05/20/introducing-rampart-and-clarity-open-source-tools-to-bring-safety-into-agent-development-workflow/) - 🆕 **May 20, 2026**. Companion to RAMPART. Structured design-review tool for AI agents — "living artifacts" documenting intent, risks, and behavior before code is written. Open-sourced from Microsoft AI Red Team's internal practice.
 - [Nobulex](https://github.com/arian-gogani/nobulex) - ⚠️ **Unverified.** Cryptographic receipts for AI agent actions (Ed25519 dual signatures, hash-chained audit logs). MIT. Bilateral-receipt primitive [merged](https://github.com/microsoft/agent-governance-toolkit/pull/1333) into Microsoft's Agent Governance Toolkit (PRs #1302, #1333). Same submission sent to 15+ awesome lists in parallel; submitter's claim of "4,500 npm downloads" doesn't match registry data (`@nobulex/mcp-server` ~19/month at audit time). Listed for visibility on the strength of the Microsoft adoption. ![GitHub stars](https://img.shields.io/github/stars/arian-gogani/nobulex?style=flat-square)
+- [ActPlane](https://github.com/eunomia-bpf/ActPlane) - OS-level agent harness that compiles a policy DSL to an eBPF engine for labeled information-flow control at the syscall boundary. Enforces constraints below the tool layer so policies hold across any tool, subprocess, or direct syscall, with corrective feedback to the agent on violation. MIT. ![GitHub stars](https://img.shields.io/github/stars/eunomia-bpf/ActPlane?style=flat-square)
 
 ## 🔍 RAG & Knowledge
 
@@ -876,6 +877,7 @@ Entries may carry one or more status tags so readers can judge maturity at a gla
 - [AutoEvals](https://github.com/braintrustdata/autoevals) - Standalone library of best-practice LLM eval scorers (factuality, JSON validity, semantic similarity, etc.) by Braintrust. Drop-in for any framework. ![GitHub stars](https://img.shields.io/github/stars/braintrustdata/autoevals?style=flat-square)
 - [BenchClaw](https://github.com/Agnuxo1/benchclaw) - ⚠️ **Unverified.** Self-described multi-dimensional agent evaluation harness (17-judge tribunal, deception detectors, 10 scoring dimensions). Repo is single-maintainer with very low independent adoption; the same submission was sent to 8+ awesome lists in parallel — one was merged at [eudk/awesome-ai-tools](https://github.com/eudk/awesome-ai-tools/pull/229), the rest are pending or declined. Listed for visibility, evaluate before relying on its scores. ![GitHub stars](https://img.shields.io/github/stars/Agnuxo1/benchclaw?style=flat-square)
 - [PromptEden](https://www.prompteden.com) - ⚠️ **Unverified.** Commercial AI-visibility monitoring service — tracks how ChatGPT, Claude, Gemini, Perplexity, Copilot, and Grok describe brands and which competitors they recommend, refreshed daily across 9+ platforms. Submitted to 10 awesome lists on the same day — promising category but listed for visibility only, evaluate before purchasing.
+- [AgentSight](https://github.com/eunomia-bpf/AgentSight) - Zero-instrumentation eBPF observability for LLM/coding agents. Captures syscall-level traces (process, file, network) without modifying the agent, enabling full-stack behavioral analysis. MIT. ![GitHub stars](https://img.shields.io/github/stars/eunomia-bpf/AgentSight?style=flat-square)
 
 ---