# Agt Security

> Provides sandboxing validation, HMAC key rotation, and budget verification to manage security boundaries under Agentic Group Theory (AGT).

- Skill: `richfrem/agt-security` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add richfrem/agt-security`
- Raw SKILL.md: https://api.skillmd.com/api/skills/richfrem/agt-security/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: richfrem (https://skillmd.com/u/richfrem)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/richfrem/agt-security

---


# Agentic Group Theory (AGT) Security Control

This skill contains tools and reference materials to manage secure local execution sandboxes, verify process hygiene limits, and rotate cryptographic bus keys.

---

## 1. Sandbox Verification

To verify that the sub-agent execution environment complies with AGT process hygiene or containerized sandboxing:
```bash
python3 plugins/cli-agents/scripts/agt_ops.py verify-sandbox
```

This performs:
1. Validating that high-risk environment variables (e.g. `ANTHROPIC_API_KEY`, `PYTHONPATH`) are scrubbed inside subprocesses.
2. Confirming that allowed path boundaries throw exceptions on out-of-bounds access.
3. Checking container status (if Docker/Podman isolation is active).

---

## 2. HMAC Key Rotation

To generate or rotate the HMAC symmetric key used to sign messages across the local control plane:
```bash
python3 plugins/cli-agents/scripts/agt_ops.py rotate-key
```

Keys are rotated dynamically and written to `${CLAUDE_PROJECT_DIR}/context/exploration/.secrets/session_hmac.key` with restricted `0600` permissions.