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

1.5 KiB

API Reference — Performing Kubernetes Penetration Testing

Libraries Used

  • subprocess: Execute kubectl commands for cluster reconnaissance and testing
  • json: Parse Kubernetes API JSON output

CLI Interface

python agent.py recon
python agent.py sa-perms [--namespace default]
python agent.py dashboards
python agent.py escape [--namespace default]

Core Functions

enumerate_cluster_info() — Cluster reconnaissance

Gathers: K8s version, node info (OS, kubelet), namespaces, services with types/ports.

test_service_account_permissions(namespace) — RBAC permission testing

Tests 8 permissions via kubectl auth can-i: get pods, list/get secrets, create pods, exec into pods, get nodes, list namespaces, create clusterroles.

scan_exposed_dashboards() — Find management interfaces

Searches for: dashboard, grafana, prometheus, kibana, jaeger, argocd, rancher, lens. Flags LoadBalancer/NodePort services as externally accessible.

check_pod_escape_vectors(namespace) — Container escape analysis

Detects: privileged mode, CAP_SYS_ADMIN/SYS_PTRACE, hostPath mounts (/, /etc, docker.sock, /proc, /sys), hostPID namespace, hostNetwork.

Dangerous Permissions (CRITICAL)

  • list secrets / get secrets --all-namespaces
  • create pods (pod creation with escalation)
  • create pods/exec (remote code execution)
  • create clusterroles (RBAC escalation)

Dependencies

System: kubectl with cluster access No Python packages required.