Controlled Response
Turn an evidence-backed decision into a controlled response — without giving VikingIntel your keys. Human approval, bounded single-use authorization, execution in your own systems, and read-only confirmation against finalized chain truth.
- 01DecideEvidence reviewed, conclusion reached by a human.
- 02ApproveA reviewer explicitly approves the bounded action.
- 03PermitSingle-use authorization, bound to one exact scope.
- 04VerifyOne atomic check: unused, unexpired, unrevoked, kill-switch clear.
- 05ExecuteRuns outside VikingIntel, under customer keys.
- 06ConfirmExecutor report checked against finalized chain state. Read-only.
VikingIntel authorizes and confirms — it never executes.
Illustrative example — the path from a decision to a confirmed response; step names are simplified.
Consequential actions wait for a person.
Nothing consequential moves without explicit human approval. Policy evaluation can recommend, restrict, or deny — but the decision to act is taken by a reviewer, not by the system. There is no autonomous execution path: approval is the gate every response passes through.
Authorization for one action — not a blank check.
An approved decision becomes a permit bound to one exact scope: the action, the source, the destination, the asset, the amount, and an expiry. The permit authorizes that action and nothing else — a different destination, a different amount, or a different permit is a different authorization entirely.
What binds a permit
- Single use — exactly one verification can claim a permit. A second attempt finds it spent and is denied.
- Scope binding — the authorization is tied to the evaluated scope; anything outside it fails verification.
- Expiry — every permit carries a deadline. An expired permit authorizes nothing; there is no grace period.
- Tamper evidence — the permit token is cryptographically bound to its scope, so a modified scope fails verification.
Revocable until the moment it is used.
A permit can be revoked any time before it is consumed. A protocol-level kill switch stops all authorization for that protocol at once — and the kill state is checked inside the same atomic verification step, so a kill landing mid-verification still denies. Approval is not execution, and authorization is not custody: VikingIntel never holds funds or private keys.
Cancel a single permit before use, with the reason recorded.
Halt all authorization for a protocol in one action.
Every approval, revocation, kill, and verification is recorded against the investigation.
Execution happens in your systems, under your keys.
VikingIntel does not submit transactions, move funds, freeze assets, or hold private keys. The authorized action is executed by your own systems or executor — VikingIntel's role ends at authorization and resumes at confirmation. What VikingIntel cannot do is as important as what it can: no custody, no autonomous execution, no guaranteed prevention.
An executor report is a claim — the chain is the truth.
After external execution, the executor reports what happened. VikingIntel treats that report as a claim and checks it against finalized on-chain state: what actually landed, in what amount, at which address. Confirmation is read-only — it changes nothing on-chain, and confirming one action never authorizes another.
The confirmation boundary
Respond with control.
Evaluate policy, approve explicitly, authorize narrowly — and confirm against chain truth.
No credit card required · Free forever