Merge improvements/fork-harvest into main

Fork-harvest brings: the ip-api.com geo fix, credential bundle scan,
mock gateways out of production, JIT compiler dropped (1.55->1.04 MB),
host hardening, provider-agnostic identity + VK/Yandex + account linking,
and the backend contracts consolidated into one BACKEND-INTEGRATION.md.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

# Conflicts:
#	docs/backend/BACKEND-HANDOFF.md
#	docs/backend/TRACK-S-SECURITY-RBAC-CONTRACT.md

scripts/deploy/server-setup.sh

@@ -128,6 +128,90 @@ ufw allow 80/tcp  >/dev/null
 ufw allow 443/tcp >/dev/null
 ufw --force enable >/dev/null
 
+# FH-D.3. ufw alone leaves SSH open to unlimited password guessing and leaves
+# the kernel on defaults that are wrong for an internet-facing host. All three
+# blocks below are drop-in files, so a re-run overwrites its own config and
+# never edits a distro file in place.
+
+echo "==> sshd hardening"
+cat > /etc/ssh/sshd_config.d/10-marketplaces-hardening.conf <<'SSHD'
+# Both accounts on this host are key-only by construction (the deploy user is
+# created with no password at all), so password auth can only ever succeed for
+# a credential nobody intended to exist.
+PasswordAuthentication no
+KbdInteractiveAuthentication no
+PermitEmptyPasswords no
+PermitRootLogin prohibit-password
+X11Forwarding no
+AllowAgentForwarding no
+MaxAuthTries 3
+LoginGraceTime 30
+ClientAliveInterval 300
+ClientAliveCountMax 2
+SSHD
+# Validate before reloading: a bad sshd config that takes effect on a remote
+# box is how people lock themselves out permanently.
+if sshd -t; then
+  systemctl reload ssh 2>/dev/null || systemctl reload sshd
+else
+  echo "sshd config test FAILED - removing the drop-in and leaving sshd as it was" >&2
+  rm -f /etc/ssh/sshd_config.d/10-marketplaces-hardening.conf
+  exit 1
+fi
+
+echo "==> fail2ban"
+apt-get install -y -qq fail2ban
+cat > /etc/fail2ban/jail.d/marketplaces.local <<'F2B'
+[DEFAULT]
+backend  = systemd
+findtime = 10m
+bantime  = 1h
+maxretry = 5
+
+[sshd]
+enabled = true
+
+[nginx-http-auth]
+enabled = true
+
+[nginx-bad-request]
+enabled = true
+F2B
+systemctl enable --now fail2ban
+systemctl restart fail2ban
+
+echo "==> kernel hardening"
+cat > /etc/sysctl.d/99-marketplaces-hardening.conf <<'SYSCTL'
+# Ignore ICMP redirects and source routing: this host has one gateway and
+# nothing upstream should be rewriting its routing table.
+net.ipv4.conf.all.accept_redirects = 0
+net.ipv4.conf.default.accept_redirects = 0
+net.ipv6.conf.all.accept_redirects = 0
+net.ipv6.conf.default.accept_redirects = 0
+net.ipv4.conf.all.send_redirects = 0
+net.ipv4.conf.default.send_redirects = 0
+net.ipv4.conf.all.accept_source_route = 0
+net.ipv6.conf.all.accept_source_route = 0
+
+# Reverse-path filtering and martian logging.
+net.ipv4.conf.all.rp_filter = 1
+net.ipv4.conf.default.rp_filter = 1
+net.ipv4.conf.all.log_martians = 1
+
+# SYN flood resistance.
+net.ipv4.tcp_syncookies = 1
+net.ipv4.tcp_max_syn_backlog = 2048
+net.ipv4.tcp_synack_retries = 2
+
+# No IP forwarding: this is a web server, not a router.
+net.ipv4.ip_forward = 0
+
+# Restrict kernel pointer and dmesg exposure to unprivileged users.
+kernel.kptr_restrict = 2
+kernel.dmesg_restrict = 1
+SYSCTL
+sysctl --quiet --system
+
 echo "==> nginx config test"
 nginx -t
 systemctl enable --now nginx