Claude 24f816b6a3
Consolidate 22 sibling repos into layered organism structure
Place useful parts of the surrounding repos into sica-fondt by layer, per the
body model (Ada = membrane; brain/endocrine/capabilities/knowledge non-Ada):

- brain/        LLM reasoning + providers (dapr, hermes, MoMoA)
- capabilities/ REPRAG sidecars: hermes tools/skills, dapr tools, parallel
                dispatch, A51 channels, and the OSINT cluster
- knowledge/    LORAG corpus: 754 cyber-skills, agency personas, secure-coding,
                MITRE ATT&CK data
- reference/    defensive threat-reference (C3, shhbruh doc) + AdaYaml parser

License handling: AGPL sources (worldosint, advanced_evolution, mercury,
Reticulum) and GPL DeTTECT are SPEC-only clean-room/port descriptions — no
copyleft code copied. MIT/Apache/data parts copied as working trees.

Safety: shhbruh escape/persistence material and C3 covert-C2 kept as reference
only, not wired into the running organism. See CONSOLIDATION.md.

https://claude.ai/code/session_01UehUqEXXJJCsHoA4voCU5c
2026-06-10 06:53:01 +00:00

67 lines
2.1 KiB
Markdown

# API Reference: Threat Hunt Hypothesis Framework
## Hypothesis Structure
| Field | Description |
|-------|------------|
| hypothesis_id | Unique identifier (HYP-XXXXXXXX) |
| technique_id | MITRE ATT&CK technique (e.g. T1059.001) |
| hypothesis_statement | Natural language hypothesis |
| data_sources | Required log sources |
| priority | high / medium / low |
| status | planned / in_progress / completed |
## MITRE ATT&CK Data Sources
```bash
# Download ATT&CK STIX bundle
curl -O https://raw.githubusercontent.com/mitre/cti/master/enterprise-attack/enterprise-attack.json
# Filter attack-pattern objects for technique data sources
python3 -c "
import json
bundle = json.load(open('enterprise-attack.json'))
for obj in bundle['objects']:
if obj.get('type') == 'attack-pattern' and not obj.get('x_mitre_deprecated'):
eid = obj['external_references'][0]['external_id']
ds = [d['source_name'] for d in obj.get('x_mitre_data_sources', [])]
print(f'{eid}: {ds}')
"
```
## Hunt Maturity Model (HMM)
| Level | Name | Description |
|-------|------|------------|
| HM0 | Initial | Ad hoc, no documented procedures |
| HM1 | Minimal | Basic procedures, limited data sources |
| HM2 | Procedural | Documented hypotheses, repeatable hunts |
| HM3 | Innovative | Custom analytics, TI-driven hypotheses |
| HM4 | Leading | Automated, ML-assisted, continuous hunting |
## Key Windows Event IDs for Hunting
| Event ID | Source | Use Case |
|----------|--------|----------|
| 4104 | PowerShell | Script block logging |
| 4688 | Security | Process creation |
| 4624/4625 | Security | Logon success/failure |
| 4698 | Security | Scheduled task created |
| 1 (Sysmon) | Sysmon | Process create with hashes |
| 3 (Sysmon) | Sysmon | Network connection |
| 10 (Sysmon) | Sysmon | Process access (LSASS) |
| 11 (Sysmon) | Sysmon | File create |
## Sigma Rule Integration
```yaml
title: Suspicious PowerShell Execution
status: experimental
logsource:
product: windows
service: powershell
detection:
selection:
EventID: 4104
ScriptBlockText|contains:
- 'Invoke-Mimikatz'
- 'Invoke-Expression'
condition: selection
level: high
```