The agentic era · governance that acts

Govern the AI agents, not just the models.

Agents plan, call tools and act. That's a new risk surface — and CSOAI was built agentic-native for it: signed agent cards, a designed 33-agent Council of AI, and Ed25519 Layer-0 attestations, mapped to EU AI Act Art. 14 & 50. Not documentation — control.

Tool misuse

An agent calls a tool it shouldn't, or with unsafe inputs.

▸ Governed MCP tools + Layer-0 signed call records
Scope drift

An agent quietly expands beyond its intended purpose.

▸ Signed agent cards declare purpose, tools & guardrails up front
Inter-agent risk

Multi-agent chains amplify errors no single agent owns.

▸ Designed council review — no single model decides; measured status on the Refutation Ledger
No human oversight

Autonomous actions with no meaningful control point.

▸ EU AI Act Art. 14 oversight gates, enforced not documented
No disclosure

Users can't tell they're dealing with AI.

▸ Art. 50 transparency — disclosure + marking at first interaction
Unaccountable actions

No provable record of what an agent did or why.

▸ Every decision sealed to Layer 0 (Ed25519), verifiable offline
Why CSOAI is a generation ahead on agents
  • Signed agent cards (A2A): every agent identified, purpose-bound and Ed25519-signed — discoverable at /.well-known/agent-card.json.
  • council review: designed so no single model approves an agent action — a designed supermajority quorum. Measured status (n_eff 1.21 of 3) is published on the Refutation Ledger — design, not yet a live claim.
  • 291 governed MCP servers: agents call tools through a governed layer, every call sealable to Layer 0.
  • Verifiable, not asserted: competitors document controls; CSOAI produces signed, reproducible proof.

Frequently asked

What is AI agent governance?

AI agent governance is the discipline of controlling autonomous AI agents — systems that plan, call tools, and act with limited human input. It covers agent identity, purpose limits, human oversight, tool-use control, inter-agent risk, and an auditable record of every action.

How is governing AI agents different from governing AI models?

A model produces an output; an agent takes actions across tools and other agents. That adds new risks — tool misuse, scope drift, and inter-agent failures — plus stronger duties for human oversight (EU AI Act Art. 14) and transparency/disclosure (Art. 50). Governance has to move from documenting a model to controlling an actor.

How does CSOAI govern AI agents?

CSOAI is agentic-native: every agent carries a signed agent card (purpose, tools, data sources, guardrails) discoverable via A2A, its actions are designed to be reviewable by a 33-agent Council of AI held to a 0.95 care-floor (a designed council — measured status is published on the public Refutation Ledger), and every decision is sealed to Layer 0 with Ed25519 for a verifiable, reproducible record — mapped to the EU AI Act, NIST AI RMF and ISO 42001.

Does CSOAI cover the EU AI Act obligations for agents?

Yes. Human-oversight duties map to Article 14, transparency/disclosure to Article 50, risk management to Article 9, and record-keeping to Articles 11–12 — all evidenced through signed attestations rather than screenshots.

Agent governance: design vs measured

n_eff 1.21 / 3
measured cross-architecture decorrelation today
measured
33 seats
target council architecture — design, not live
design
0
mechanisms sold as live that are design-stage
measured

source: Refutation Ledger DR-0007 + gate1 decorrelation runs, 2026-08-01

FAQs

Frequently asked questions

How governed agents are measured — design vs measured, always labelled.

What is agent governance in measurable terms?

Each governed action is decomposed into checks we can run and count: who proposed it, what policy applied, whether the care floor held, and whether the record is signed. If a check cannot run, it is reported UNMEASURED — not assumed.

Does the council approve agent actions live?

That is the design, not a live claim. The measured cross-checking status today is n_eff 1.21 of 3, published on the Refutation Ledger. The 33-seat council architecture is labelled as a design simulation wherever it is shown.

What stops a single model from approving its own action?

The design rule is that no single model approves an action — a supermajority quorum does. What we can evidence today is the measured decorrelation between independent architectures; the full quorum is published as a target, not a result.

Where are the refuted claims?

On the public Refutation Ledger. When a measurement shows a mechanism does not deliver what the design predicted, the refutation is published with its n and confidence interval — the same prominence as a success.