CyberHalluciNet CHN
Install Phase A+C

Agentic platform

Core engine architecture

Four specialized roles across hard trust boundaries. Three deception surfaces. Closed-loop containment that stays fail-closed, models never own the wire.

Roles Surfaces Safety Deploy

Authority matrix

The 4 specialized agent roles

Who may touch what. Wire authority stays in the sensor; AI stays suggestion-only.

01 Wire plane

Sensor

Deterministic wire-level execution. Policy, personas, and synthetic responses: score-independent, fail-closed.

  • Owns attacker-visible bytes
  • No model chooses egress / OS exec
02 Suggestion plane

Content Agent

Bounded, schema-validated AI fills over UDS. Fail-closed when the broker is unavailable.

  • ValidateSlotFill gate
  • Static / template fallback
03 Control plane

Fabric Planner & Scenario Director

Operator-approved deception change management with a signed audit ledger.

  • Propose → Approve → Deploy
  • Never on the attacker wire
04 Analyze plane

Read-Only CTI Investigator

Cited evidence assembly without mid-session risk or auto-containment.

  • Closed read-only tools
  • No wire morph from scores

Attacker path

Attacker Sensor Content Agent
UDS · slot fills

Operator path

Operator Fabric Planner CTI Investigator
read-only · cited

Planes stay separated. Scores never grant wire authority.

Deception fabric

The 3-surface deception fabric

Purpose-built coverage for enterprise identity, cyber-physical plants, and AI agent infrastructure.

Identity & host defense

Identity lures (ITDR-style)

ITDR = Identity Threat Detection and Response. Agentless host and directory-oriented tripwires that end at honeypot surfaces : without a heavyweight endpoint agent as the default path.

  • Registry / SSH config lure packs (GPO / Intune / Jamf-oriented)
  • Monitored AD / Entra-style honey account patterns
  • Self-cleaning TTL lifecycle to cut orphaned debt

Built for · CISO & IT SecOps

Safety interlocks

Closed-loop containment

Human-gated, decoy-touch scoped. No score-driven IPS. No silent cloud mutation.

Pre-flight blast-radius

Simulate exposure before permissive binds or Live/ack-gated paths. Sign-off before the blast radius grows.

Rate-limit circuit breakers

Admission and sticky per-source controls keep CTI / Internet postures from runaway scanner floods.

Immutable exclusion matrices

Hard exclusions keep protected ranges, production IAM, and non-decoy assets off the auto path.

Phase A wired (defaults off) · Phase C Beta (LocalState + BYO webhook)

Agentic attack defense

Deterministic decoy-touch evidence for compromised AI agents. Complements guardrails; does not replace them. Soft/Hard are Beta via LocalState and optional customer webhook, native vendor EDR/IdP SDKs remain GA-train. Canonical wording lives in the docs positioning page: this section must not exceed it.

Install in a lab: Install Phase A · Install Phase C Beta · All agentic tutorials (make agentic-lab-install).

Graded evidence

observed_anomaly → verified_decoy_touch / canary_egress / credential_use. Scores never promote a grade.

Do-no-harm retrieval

Node-local shadow index for canary memory, production top-k stays clean.

Bounded containment (Phase C Beta)

Separate Enforcement Plane; session-first ladder; rung 3 never autonomous. Soft/Hard via LocalState + BYO webhook. Native vendor APIs are GA-train only.

Non-claim: third-party canary benchmarks describe those studies, not CHN performance. Internal use remains free under PSL-1.0; no online entitlement check for enforcement.

What is CyberHalluciNet agentic attack defense?

Deception that fires when a tool-calling agent or operator touches honeytools, shadow canaries, or other bait. Phase A is advisory with defaults off. Phase C Beta Soft/Hard runs on a separate Enforcement Plane via LocalState and optional BYO webhook; native vendor EDR/IdP SDKs remain GA-train only.

Why keep containment off the sensor?

The sensor notify path is evaluate-only. Ops holds the execute token and connector credentials stay on the Enforcement Plane, so a compromised sensor identity cannot Soft/Hard by itself.

How do I install the Phase A and Phase C Beta lab?

Run make agentic-lab-install from a repo checkout, then follow the Install Phase A and Install Phase C Beta tutorials. Confirm /healthz reports maturity beta before Soft/Hard pilots.

Next step

Ready to test?

Deploy a staging lab in about ten minutes, tutorials walk the path without touching production.

Ready to Test? Deploy in 10 Minutes Explore The A.I. Decoy Engine