mirror of
https://github.com/SHOGGOTH-SECTOR/sica-fondt.git
synced 2026-08-01 08:30:20 +00:00
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
1.9 KiB
1.9 KiB
Falco Container Escape Detection Runbook
Alert Triage Template
Alert Details
| Field | Value |
|---|---|
| Alert Time | |
| Rule Name | |
| Priority | |
| Container Name | |
| Container Image | |
| Pod Name | |
| Namespace | |
| Node | |
| User | |
| Process Command | |
| MITRE Technique |
Triage Steps
Immediate Actions (0-5 minutes)
- Acknowledge alert in SIEM/SOAR
- Verify alert is not a false positive (check known exceptions list)
- Identify the affected pod and node
- Check if the container is still running
Investigation (5-30 minutes)
- Capture pod spec:
kubectl get pod <name> -n <ns> -o yaml - Review container security context
- Check if container is privileged
- Review mounted volumes for host paths
- Examine process tree from Falco output
- Check for other alerts from same container/node
- Review Kubernetes audit logs for the same timeframe
Containment (if confirmed)
- Isolate pod with network policy deny-all
- Cordon affected node:
kubectl cordon <node> - Capture forensic data from container
- Kill compromised container:
kubectl delete pod <name> -n <ns> - Review other pods on same node for compromise
Recovery
- Scan node for rootkits
- Rebuild node if compromise confirmed
- Patch vulnerable container image
- Update network policies
- Uncordon node after verification
False Positive Exceptions
| Container Image | Rule | Justification | Approved By | Date |
|---|---|---|---|---|
Escalation Matrix
| Priority | Response Time | Notify |
|---|---|---|
| CRITICAL | Immediate | Security On-Call + Engineering Lead |
| WARNING | 15 minutes | Security On-Call |
| NOTICE | 1 hour | Security Team queue |
| INFO | Next business day | Review in daily standup |