Atested sits in front of governed AI operations. Every action is evaluated against the same logic a well-informed engineer would apply — then the evaluation and the decision are signed and provable.
The core flow is simple. An AI tool routes an action through Atested. Atested evaluates it against verifiable conditions — the scope, evidence, and constraints that any well-informed engineer would require. If the conditions are met, the action proceeds and the decision is recorded. If not, the action is denied and that denial is recorded too.
{
"tool": "fs_write",
"capability_class": "FS_WRITE",
"policy_decision": "ALLOW",
"timestamp_utc": "2026-03-30T13:12:00Z",
"operator_intent": "update README",
"organization_id": "acme-engineering",
"record_hash": "sha256:0a1b2c3d4e5f...",
"prev_record_hash": "sha256:f0e1d2c3b4a5...",
"signature": "ed25519:8f9c2ab1..."
}
Atested does not just log actions after the fact. It evaluates governed actions before they proceed. ALLOW means the action met the verifiable conditions — the evidence was sufficient, the scope was valid, the constraints were satisfied. DENY means something was missing, and the record shows exactly what.
A coding agent writes inside an approved project path using an allowed tool and the required execution context. An agent sends a message through an approved channel with the expected recipient scope and supporting context.
An agent tries to edit files outside the approved workspace, touch production configuration without authority, or assert that tests passed without sufficient evidence to support the claim.
Atested governs every action that flows through it. It cannot force every action to flow through it, because AI tools also have native capabilities that operate outside any governance layer. That is a structural reality in open environments.
What Atested does is make that boundary visible and measurable. Governed operations produce full signed records. Ungoverned native operations can still be observed and counted through action hooks, so you can see how much of your AI activity is actually under governance.
{
"governed_operations": 1842,
"observed_native_operations": 716,
"transparency_ratio": "72%",
"observation_mode": "hook-reported",
"manager_view": "governed vs observed"
}
Governance decisions fall into two kinds: those that can be determined logically, and those that require judgment. Atested resolves as many as possible deterministically — is the evidence present, is the scope valid, are the preconditions met. These aren't arbitrary rules. They're the same questions a careful engineer would ask if they were reviewing every action by hand. Atested asks them consistently, every time, and signs the answer.
The conditions aren't invented. They're logical necessities. If an AI agent wants to write to a file, the deterministic questions are: does the file path resolve to a real, approved location? Is the agent authenticated? Is there a valid chain of decisions leading to this action? A well-informed engineer would ask exactly the same things.
Where a decision can't be resolved deterministically, Atested says so. Scoped approvals exist for exactly this case — you make the judgment call, and that judgment is recorded alongside the deterministic decisions. The system is honest about what it can prove and what required your judgment.
Change the approved artifact and the approval expires. Granting and revoking approvals are themselves recorded in the governance chain, so exceptions stay visible and auditable.
The Atested Dashboard gives your team a live view of governance activity, decisions, approvals, audit results, and system health — all backed by the same signed chain the governance engine produces.
No. Atested's governance logic reflects the conditions any well-informed engineer would apply — is the evidence present, is the scope valid, are the preconditions met. You configure scope and constraints for your environment, but the evaluation logic is built in. Most decisions are resolved deterministically without any rules to write.
Without a governance layer, AI operations often become reconstruction work after something goes wrong. With Atested, many unsupported actions can be denied before they land, and the resulting record gives you and your team a usable account of what happened instead of a black box and a pile of logs.