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

2.8 KiB

Docker Container Hardening Assessment Template

Project Information

Field Value
Application Name
Docker Image
Base Image
Environment Development / Staging / Production
Assessment Date
Assessor

Pre-Hardening Checklist

Dockerfile Security

  • Using minimal base image (distroless, Alpine, scratch)
  • Specific image tag with digest pinning (not :latest)
  • Multi-stage build implemented
  • Non-root USER instruction present
  • COPY used instead of ADD
  • No secrets in Dockerfile or image layers
  • HEALTHCHECK instruction present
  • Unnecessary packages removed
  • setuid/setgid binaries removed

Daemon Configuration (/etc/docker/daemon.json)

  • icc set to false
  • TLS authentication enabled (tlsverify: true)
  • Live restore enabled
  • Userland proxy disabled
  • no-new-privileges enabled
  • Log rotation configured (max-size, max-file)
  • Default ulimits configured
  • Seccomp profile specified

Runtime Security Flags

  • --read-only enabled
  • --cap-drop ALL applied
  • Minimum --cap-add for required capabilities only
  • --security-opt no-new-privileges:true
  • --security-opt seccomp=
  • --memory limit set
  • --cpus limit set
  • --pids-limit set
  • --user set to non-root UID:GID
  • --tmpfs for writable directories
  • --network set to custom bridge (not host)
  • --restart on-failure with max retries

Host Security

  • Separate partition for /var/lib/docker
  • Docker group membership restricted
  • Audit rules configured for Docker files
  • Docker socket not exposed to containers
  • Content Trust enabled (DOCKER_CONTENT_TRUST=1)

Vulnerability Scan Results

Trivy Scan

trivy image <image-name>
Severity Count Action Required
CRITICAL Immediate fix
HIGH Fix before production
MEDIUM Plan remediation
LOW Accept or fix

Docker Bench Score

docker run --rm docker/docker-bench-security
Section Score Notes
Host Configuration /10
Daemon Configuration /10
Container Images /10
Container Runtime /10
Docker Security Ops /10

Risk Acceptance

Finding Severity Justification Approved By Date

Remediation Plan

Priority Finding Action Owner Target Date Status
P1
P2
P3

Sign-Off

Role Name Signature Date
Security Engineer
DevOps Lead
Application Owner