Sigilant · Agent identity gateway
Every agent action, authorised.
Sigilant sits between your AI agents and the systems they act on. Every agent has its own identity and an accountable owner, every action is checked before it happens, and every decision is on the record.
The gap
Agents act with authority nobody gave them.
An agent that can reach a system usually reaches it with a shared key or a person's standing access. That settles whether it can act. It does not settle whether it should, on whose behalf, or who signed off.
Who
No identity of its own
Agents borrow service accounts and API keys. In the logs, ten agents and one integration look the same.
What
Access, not authorisation
A credential says what a system will accept. It says nothing about whether this action, for this person, at this moment, is one you would allow.
Proof
Evidence after the fact
When an auditor or an incident asks what an agent did and who approved it, the answer is assembled by hand, if it can be assembled at all.
How it works
One checkpoint. Four things happen at it.
Agents call their tools through Sigilant instead of directly. Neither the agent nor the system behind it has to be rebuilt.
Step 01
Decide
Each action is checked against your policy before it runs: which agent, for whom, doing what.
Allow, hold or refuse
Step 02
Approve
Actions you mark as high-risk are held for a named person, who sees exactly what they are approving.
A person in the loop
Step 03
Credential
The agent never holds the keys to your systems. Access is supplied for the one action, then gone.
Nothing to steal
Step 04
Record
Every decision, allowed or refused, is written to a record that shows if it has been altered.
Evidence as a by-product
Capabilities
What you get.
Governance for agents, with the rigour you already expect for people.
Identity
An identity for every agent
Each agent is registered, has an accountable owner, and is distinguishable from every other in every record.
Delegation
Authority that only narrows
When one agent hands work to another, the second can never do more than the first was allowed to.
Control
A check on every action
Policy is applied to each individual action as it happens, not once when access is first granted.
Oversight
Approval where it matters
Routine work flows. Actions over your thresholds wait for the right person, and the approval covers that action only.
Response
A switch that works at once
Stop one agent, one task, or everything an agent has started. It takes effect immediately, including on work in flight.
Assurance
A record you can verify
Audit evidence is produced as the work happens and can be checked independently, by you or by your auditor.
Deployment
Built to fit what you already run.
Sigilant works alongside your existing identity provider, so people sign in and approve the way they already do. Ownership of agents follows the people and groups you already manage.
It runs in your environment. Your policies, your records and the detail of what your agents do stay with you.
Sigilant is in early access. We are happy to walk through it against your own agents and systems.
Know what your agents did, and who said they could.
Bring one agent and one system it touches, and we will show you what governing that path looks like.
Get in touch
Request a briefing.
Tell us which agents you are putting into production and what they need to reach. We will reply within one business day.
- Email[email protected]
- Phone1300 786 040
- LinkedInCintelis AI
- BasedBrisbane, Queensland