Full Constitutional Stack — All 11 Agents

The Full Constitutional Stack — governance authority, execution power, operational visibility. All 10 agents.

The most hardened autonomous AI stack available for OpenClaw.

The Full Constitutional Stack is the complete HiveDeck dual-authority architecture — all three tiers, all 10 agents, one install sequence. Governance runs first. Execution operates under it. Operations supports everything. This is what a production-grade autonomous agent system looks like when it's built right.

One install-all.sh. Every agent. Every safeguard. The full constitutional AI pipeline.

All 10 agents. Every tier. Complete architecture.

Tier 1 — Governance (3 agents)

• 🛡️ Sentinel — Veto & Enforcement Authority. The final word on what runs. Every agent action reviewed, dangerous actions stopped before execution.
• 🔒 Shield — Security Hardening. Hardens your environment, enforces least-privilege access, scans for vulnerabilities before execution agents run.
• 📋 Auditor — Compliance & Validation. Validates every output before it propagates. GIGO enforcement. The agent that catches mistakes.

Tier 2 — Execution (5 agents)

• 🔍 Scout — Research & Intelligence. Deep competitive analysis, market research, technology evaluations. Stop Googling.
• ⚒️ Forge — Implementation. Writes code, deploys services, runs builds. The agent that ships.
• 📊 Ranker — Evaluation & Prioritization. Scores options, breaks ties, surfaces clear recommendations.
• 🔷 Prism — Decomposition. Breaks complex objectives into clean, executable task sequences.
• 🤝 Closer — Completion. Monitors open loops, drives stalled work to done.

Tier 3 — Operations (2 agents)

• 📒 Ledger — Resource Tracking. Every API dollar accounted for. Daily cost reports. No budget surprises.
• ✍️ Scribe — Documentation. Changelogs, runbooks, decision logs — generated automatically from agent activity.

What you get: one install, complete infrastructure

• 10 complete agent packages — SOUL.md, skills/, TOOLS.md, install/uninstall/verify for every agent
• install-all.sh — deploys all 10 in the correct dependency order (governance first, then execution, then operations)
• Full integration guide — complete AGENTS.md wiring instructions for all three tiers
• ClawHavoc safety protocols — the full constitutional ruleset, incident response, boundary definitions
• Inter-agent coordination configs — pre-wired handoff and communication protocols across all 10 agents
• Dashboard templates — cost and documentation reporting, ready to run

Install in 4 steps

1. Download the ZIP
2. cd ~/.openclaw/agents && unzip full-constitutional-stack.zip
3. ./install-all.sh
4. Tell your Chief of Staff: "Full constitutional architecture is active — all tiers online"

Works with Claude 3.5/4 or GPT-5. This is the complete system. Every governance control, every execution capability, every operational safeguard — deployed together, working as designed.

This is what constitutional AI infrastructure looks like.

Not a chatbot. Not a prompt library. A dual-authority architecture where governance controls execution, execution operates under oversight, and operations keeps everything measured and documented. 10 agents. Three tiers. One coherent system. Built for OpenClaw users who are serious about running autonomous AI at production grade.

Best value in the HiveDeck catalog. Save vs. buying all stacks separately. One install-all.sh deploys everything.

New to HiveDeck? Step-by-step install guide →