Security agents
Agents that investigate alerts and audit configuration across the estate — read-only unless a person decides otherwise.
Investigate this alert across the log platform, the endpoint tool and the host.
A per-host account of what was checked and what was found, with the exact commands behind it.
The use case, on film
Named agents by vertical
What a security agent does
An investigation is a sequence of questions, each one decided by the answer before it. The analyst who is good at it is good at knowing what to look at next, and spends most of the shift collecting rather than deciding. The collection is the part that can be handed over — provided nothing is changed by collecting it.
- Follow an alert from the log platform to the endpoint tool to the machine.
- Audit one setting across the whole estate and answer per machine.
- Report the branches that were tried and found empty, not only the ones that led somewhere.
- Stop before containment and wait for the person who owns the decision.
The collection those bullets describe is what an agent carries out on the hosts themselves:
Host investigations show the commands that ran, their exit codes and output, per host.
The hosts and skills behind it
Host Advisor inspects enrolled machines and returns what it ran; the named investigators around it are agent objects composed on top of that same execution and evidence.
Read-only execution policy is the default on every enrolled host; the full policy is an explicit, per-host choice.
A host joins the fleet with one install command and an enrollment token; revoking the host unpairs it.
Skills ship with the release, come from AcTree's signed skill catalogue, or are your own uploads, and each workspace enables or disables each one.
The primitives it composes
An agent is a model, a set of approved skills, tools and a policy, applied to a scope.
Each skill is a versioned package. Uploads are checked before they install, every version is archived, and each workspace enables or disables each skill.
Agents reach your systems through the MCP servers you connect in the console. A token you enter there is sealed, never shown again, and acts only with the permissions your system gives it; a connector that runs on one of your hosts keeps its token in that host's own connector file.
Execution follows the configured host policy. A command can be approved, denied or referred for human review; individual approval depends on that policy.
Bring one question of your own.