YAML adversary simulation profiles used by the apt_playbooks module and the
autonomous daemon. Each file maps a named threat actor or malware family to a
sequence of MITRE ATT&CK techniques backed by Atomic Red Team test IDs.
Distinct from playbooks/ in scope: files here are authoritative threat actor
profiles (static reference data). Files in playbooks/ are engagement-specific
attack plans generated or customised per target.
| File | Threat actor | Primary MITRE tactics |
|---|---|---|
apt_apt28.yaml |
APT28 / Fancy Bear (GRU) | Credential access, lateral movement, exfiltration |
apt_apt29.yaml |
APT29 / Cozy Bear (SVR) | Persistence, defense evasion, C2 |
apt_apt41.yaml |
APT41 (dual espionage + financial) | Initial access, execution, privilege escalation |
apt_conti.yaml |
Conti ransomware group | Execution, lateral movement, impact |
apt_fin7.yaml |
FIN7 / Carbanak | Spearphishing, credential access, collection |
apt_lazarus.yaml |
Lazarus Group (North Korea) | Supply chain, financial theft, destructive ops |
apt_lockbit.yaml |
LockBit ransomware | Lateral movement, data encryption, extortion |
apt_name: apt28
description: "APT28 (Fancy Bear) — GRU Unit 26165 TTPs based on published CTI"
mitre_groups: ["G0007"]
references:
- "https://attack.mitre.org/groups/G0007/"
phases:
- phase: credential_access
techniques:
- technique_id: T1003.001
name: LSASS Memory Dump
atomic_ids: ["T1003.001-1", "T1003.001-2"]
platforms: ["windows"]
detection_notes: "Monitor for lsass.exe handle access with PROCESS_VM_READ"During an engagement, the daemon reads the active adversary profile from
sessions/world_model.json (current_adversary field). It filters techniques
by the target OS (os_id from payload.json) and selects the next executable
technique based on the current kill-chain phase.
To set the adversary simulation profile:
(LazyOwn) > assign current_adversary apt29
Or via MCP:
lazyown_set_config(key="current_adversary", value="apt29")
- Create
lazyadversaries/apt_<name>.yamlfollowing the structure above. - Add
mitre_groupsIDs for attribution links. - Map each technique to at least one Atomic Red Team test ID where available.
- Mark platform-specific techniques explicitly — the daemon skips Windows techniques against Linux targets.
- Run
python scripts/update_apt_atomic_ids.pyto resolve and cache Atomic test metadata.