A field guide to agent-system risk
A permission tells you what one agent can do. Aegis Relay reveals what a whole network can become.
Trace the route from routine actions to an unintended capability. Review the evidence, compare the available controls, and verify what remains.
Fictional environments. Deterministic analysis. Simulated controls.
01Observe the relationshipsBuild the system view from reported interactions.
02Challenge the compositionCompare observed paths with the declared architecture.
03Keep the intended workVerify the result after a registered control is applied.
The moving system
The sequence below is an illustration. Open the workspace for replayable output from the analysis engine.
Evaluated one at a time, every actor is compliant. Each check is correct — and each check is blind to the others.
Connect the same three actors through shared state and delegation and a capability appears that nobody granted: read confidential data, move it, send it out.
A shared resource can connect two otherwise separate jobs. Follow those connections to see when the system gains a capability its architecture never declared.
Behavior produced by interactions between individually valid components that was never explicitly designed.
Aegis Relay does not replace agents, tools or identity systems. It observes the ecosystem around them and models what the population is doing as a whole.
Observation plane · not an execution dependencyChange the conditions. Follow the result.
Set a population, vary its permissions and shared state, and introduce a perturbation. This conceptual lab illustrates how a new route can form; the workspace runs the implemented engine over its generated environments.
{{ findBody }}
Aegis Relay doesn't only ask whether an agent is compromised. It asks whether the system itself is entering a state it was never designed to reach.
Aegis Relay continuously compares observed population behavior against the system's expected behavioral baseline across seven signal families. Deviation is the signal — not a malware match.
The Emergence Score measures how far observed population behavior has deviated from the system's expected behavioral baseline. It is not a probability of attack, and the values shown here are illustrative of the concept rather than a validated production measurement.
Agents, tools, data, memory, MCP servers, identities and external systems — one graph, continuously rebuilt from observed events.
Compare the options before you interrupt.
Compare the response options against the same network. Preview each candidate, inspect the relationships it would interrupt, and keep the work the architecture intends.
{{ ivNote }}
Impact figures are computed inside this simulated environment from the graph shown above. They illustrate how a containment planner reasons about blast radius — they are not measured results from a production deployment.
A resolved path leaves a useful structural pattern. Capture that pattern for review, then test whether it holds in another environment before promoting it.
{{ immRule }}
A stored pattern begins as a draft. Re-simulation helps evaluate it; a person decides whether it moves to shadow or active use.
Securing a population is continuous: what Swarm Lab discovers becomes what Response planning watches for, and what production reveals becomes the next simulation.
{{ lcBody }}
{{ lcMeta }}Discover dangerous behaviors before production — by running the population, not the agent.
synthetic population · controlled perturbationDetect and contain dangerous system states as they emerge — with the smallest possible intervention.
live topology · minimal containmentTurn discoveries into future resilience — as reusable structure, not a single indicator.
pattern corpus · candidate immunityKeep the controls that protect models, prompts, identities, tools and data. Add a system view to inspect the capabilities that their interactions assemble.
A support agent receives a malicious instruction inside retrieved content. The story is not the text — it is what the population does next. No jailbreak, no stolen credential, no compromised model: every actor stays inside its own policy.
{{ thBody }}
This illustrative stream runs in your browser. The investigation workspace separately replays recorded engine output or connects to your local analysis service.
As autonomous systems become more distributed, the number of possible interactions grows faster than a human operator can reason about manually.
Pairwise combinations only — not observed interactions. Edges sampled for display.
Every one of those multiplies the states a system can reach — and a state nobody designed is a state nobody reviewed.
Explore the research behind system-level agent security. These independent sources frame the problem; they do not endorse or validate Aegis Relay.