Integration proposal: Agentegrity — agent integrity and security verification for ADK

**Agentegrity** is an open agent-integrity and security-verification framework from Cogensec with native Google ADK instrumentation.

It adds:

- Agent integrity scoring
- Adversarial-input and prompt-injection detection
- Behavioral drift monitoring
- Reasoning and memory-integrity checks
- Decision provenance
- Cryptographically signed attestations
- Tamper-evident audit chains
- Recovery verification

Agentegrity instruments ADK agents without replacing the ADK runtime or changing the agent architecture.

**Repository:** https://github.com/Cogensec/agentegrity

**Proposed catalog category:** Observability / Security

**Proposed catalog description:**

> Verify ADK agent integrity with adversarial detection, behavioral drift monitoring, signed attestations, decision provenance, and security evidence.

The integration is already implemented in the published Agentegrity Python package through its Google ADK adapter.

Installation:

```bash
pip install "agentegrity[google-adk]"
```

Basic usage:

```python
import agentegrity

runtime = agentegrity.init()

secured_agent = runtime.instrument(agent)

# Run the ADK agent normally.

report = runtime.report()

agentegrity.shutdown()
```

Review the issue proposal and the linked Cogensec/agentegrity repository, focusing on the published Python package's Google ADK adapter and the requested catalog category and description. Confirm the project's process and acceptance criteria for external integrations; done would be a maintainer-approved catalog entry or a documented decision on the proposal.

Written by the indexing model from the issue text.