You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Lab Grafana and BloodHound answer all of VLAN 30; scratch re-exposes the SOC UIs #1139 fenced #1428
Traffic between VLAN 30 hosts never crosses pfSense, so a port published on a VLAN 30 guest answers the whole segment, including the six deliberately weak domain guests. #1139 fenced odin's SOC UIs behind a Hicks-only Caddy for exactly this reason. These three were left open:
alexander's lab Grafana (3000), published by its own container.
eden's BloodHound (8443), which is the domain's attack-path graph, published by its own container.
stacks/scratch (diabolos, the detonation guest, not built yet) is "soc's shape" but publishes the Wazuh dashboard and Velociraptor GUI on 0.0.0.0, with no DIFFERENCE note.
A red-team exercise from the weak domain could reach the defenders' tools by accident, and a detonated sample could reach the SIEM's and the DFIR tool's login pages.
Evidence
stacks/lab/compose.yaml:343, stacks/bloodhound/compose.yaml:203, stacks/scratch/compose.yaml:237,284 (on 4a0c2d2); soc's fence in stacks/soc/Caddyfile and compose.yaml:390-396, 478-480.
a Caddy publishes the same host port, terminates TLS with the UI's own leaf, verifies the upstream against the CA, answers remote_ip 10.0.50.0/24, and returns 403 to everyone else
Lab adds the site to its existing Caddy, so the existing Caddyfile check validates it. BloodHound and scratch get their own Caddy.
Out of scope here: the lab ingest proxy's plain HTTP (the other half of audit finding #9). It needs a server certificate and a trust change on six agents, so it gets its own issue and PR.
Source: Phase 1 audit, finding #9 (owner: fence before the first exercise).
Activity