fix: feat: merge chat container into edge — run chat server inside edge container with full permissions (reverts sandbox from #706) (#1083)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

docker-compose.yml

@@ -208,8 +208,8 @@ services:
 
   edge:
     build:
-      context: docker/edge
+      context: .
-      dockerfile: Dockerfile
+      dockerfile: docker/edge/Dockerfile
     image: disinto/edge:latest
     container_name: disinto-edge
     security_opt:
@@ -220,6 +220,8 @@ services:
       - ${CLAUDE_CONFIG_FILE:-${HOME}/.claude.json}:/root/.claude.json:ro
       - ${CLAUDE_DIR:-${HOME}/.claude}:/root/.claude:ro
       - disinto-logs:/opt/disinto-logs
+      # Chat history persistence (merged from chat container, #1083)
+      - ${CHAT_HISTORY_DIR:-./state/chat-history}:/var/lib/chat/history
     environment:
       - FORGE_SUPERVISOR_TOKEN=${FORGE_SUPERVISOR_TOKEN:-}
       - ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY:-}
@@ -231,6 +233,16 @@ services:
       - PRIMARY_BRANCH=main
       - DISINTO_CONTAINER=1
       - FORGE_ADMIN_USERS=disinto-admin,vault-bot,admin
+      # Chat env vars (merged from chat container into edge, #1083)
+      - CHAT_HOST=127.0.0.1
+      - CHAT_PORT=8080
+      - CHAT_OAUTH_CLIENT_ID=${CHAT_OAUTH_CLIENT_ID:-}
+      - CHAT_OAUTH_CLIENT_SECRET=${CHAT_OAUTH_CLIENT_SECRET:-}
+      - DISINTO_CHAT_ALLOWED_USERS=${DISINTO_CHAT_ALLOWED_USERS:-}
+      - FORWARD_AUTH_SECRET=${FORWARD_AUTH_SECRET:-}
+      - EDGE_TUNNEL_FQDN=${EDGE_TUNNEL_FQDN:-}
+      - EDGE_TUNNEL_FQDN_CHAT=${EDGE_TUNNEL_FQDN_CHAT:-}
+      - EDGE_ROUTING_MODE=${EDGE_ROUTING_MODE:-subpath}
     ports:
       - "80:80"
       - "443:443"

docker/chat/Dockerfile

@@ -1,37 +0,0 @@
-# disinto-chat — minimal HTTP backend for Claude chat UI
-#
-# Small Debian slim base with Python runtime and Node.js.
-# Chosen for simplicity and small image size (~100MB).
-#
-# Image size: ~100MB (well under the 200MB ceiling)
-#
-# Claude CLI is baked into the image — same pattern as the agents container.
-
-FROM debian:bookworm-slim
-
-# Install Node.js (required for Claude CLI) and Python
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    nodejs npm python3 \
-    && rm -rf /var/lib/apt/lists/*
-
-# Install Claude Code CLI — chat backend runtime
-RUN npm install -g @anthropic-ai/claude-code@2.1.84
-
-# Non-root user — fixed UID 10001 for sandbox hardening (#706)
-RUN useradd -m -u 10001 -s /bin/bash chat
-
-# Copy application files
-COPY server.py /usr/local/bin/server.py
-COPY entrypoint-chat.sh /entrypoint-chat.sh
-COPY ui/ /var/chat/ui/
-
-RUN chmod +x /entrypoint-chat.sh /usr/local/bin/server.py
-
-USER chat
-WORKDIR /var/chat
-
-EXPOSE 8080
-HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
-  CMD python3 -c "import urllib.request; urllib.request.urlopen('http://localhost:8080/health')" || exit 1
-
-ENTRYPOINT ["/entrypoint-chat.sh"]

docker/chat/entrypoint-chat.sh

@@ -1,37 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-# entrypoint-chat.sh — Start the disinto-chat backend server
-#
-# Exec-replace pattern: this script is the container entrypoint and runs
-# the server directly (no wrapper needed). Logs to stdout for docker logs.
-
-LOGFILE="/tmp/chat.log"
-
-log() {
-    printf '[%s] %s\n' "$(date -u '+%Y-%m-%d %H:%M:%S UTC')" "$*" | tee -a "$LOGFILE"
-}
-
-# Sandbox sanity checks (#706) — fail fast if isolation is broken
-if [ -e /var/run/docker.sock ]; then
-    log "FATAL: /var/run/docker.sock is accessible — sandbox violation"
-    exit 1
-fi
-if [ "$(id -u)" = "0" ]; then
-    log "FATAL: running as root (uid 0) — sandbox violation"
-    exit 1
-fi
-
-# Verify Claude CLI is available (expected via volume mount from host).
-if ! command -v claude &>/dev/null; then
-    log "FATAL: claude CLI not found in PATH"
-    log "Mount the host binary into the container, e.g.:"
-    log "  volumes:"
-    log "    - /usr/local/bin/claude:/usr/local/bin/claude:ro"
-    exit 1
-fi
-log "Claude CLI: $(claude --version 2>&1 || true)"
-
-# Start the Python server (exec-replace so signals propagate correctly)
-log "Starting disinto-chat server on port 8080..."
-exec python3 /usr/local/bin/server.py

docker/chat/server.py

@@ -41,7 +41,7 @@ import base64
 import hashlib
 
 # Configuration
-HOST = os.environ.get("CHAT_HOST", "0.0.0.0")
+HOST = os.environ.get("CHAT_HOST", "127.0.0.1")
 PORT = int(os.environ.get("CHAT_PORT", 8080))
 UI_DIR = "/var/chat/ui"
 STATIC_DIR = os.path.join(UI_DIR, "static")

docker/edge/Dockerfile

@@ -1,6 +1,12 @@
 FROM caddy:latest
-RUN apk add --no-cache bash jq curl git docker-cli python3 openssh-client autossh
+RUN apk add --no-cache bash jq curl git docker-cli python3 openssh-client autossh \
-COPY entrypoint-edge.sh /usr/local/bin/entrypoint-edge.sh
+    nodejs npm
+# Claude Code CLI — chat backend runtime (merged from docker/chat, #1083)
+RUN npm install -g @anthropic-ai/claude-code@2.1.84
+COPY docker/edge/entrypoint-edge.sh /usr/local/bin/entrypoint-edge.sh
+# Chat server and UI (merged from docker/chat into edge, #1083)
+COPY docker/chat/server.py /usr/local/bin/chat-server.py
+COPY docker/chat/ui/ /var/chat/ui/
 
 VOLUME /data
 

docker/edge/entrypoint-edge.sh

@@ -244,6 +244,9 @@ else
   echo "edge: collect-engagement cron skipped (EDGE_ENGAGEMENT_READY=0)" >&2
 fi
 
+# Start chat server in background (#1083 — merged from docker/chat into edge)
+(python3 /usr/local/bin/chat-server.py 2>&1 | tee -a /opt/disinto-logs/chat.log) &
+
 # Nomad template renders Caddyfile to /local/Caddyfile via service discovery;
 # copy it into the expected location if present (compose uses the mounted path).
 if [ -f /local/Caddyfile ]; then

lib/generators.sh

@@ -326,7 +326,6 @@ _generate_compose_impl() {
   _record_service "edge" "base compose template" || return 1
   _record_service "staging" "base compose template" || return 1
   _record_service "staging-deploy" "base compose template" || return 1
-  _record_service "chat" "base compose template" || return 1
 
   # Extract primary woodpecker_repo_id from project TOML files
   local wp_repo_id
@@ -615,6 +614,16 @@ COMPOSEEOF
       - EDGE_TUNNEL_FQDN_CHAT=${EDGE_TUNNEL_FQDN_CHAT:-}
       # Shared secret for Caddy ↔ chat forward_auth (#709)
       - FORWARD_AUTH_SECRET=${FORWARD_AUTH_SECRET:-}
+      # Chat env vars (merged from chat container into edge, #1083)
+      - CHAT_HOST=127.0.0.1
+      - CHAT_PORT=8080
+      - CHAT_OAUTH_CLIENT_ID=${CHAT_OAUTH_CLIENT_ID:-}
+      - CHAT_OAUTH_CLIENT_SECRET=${CHAT_OAUTH_CLIENT_SECRET:-}
+      - DISINTO_CHAT_ALLOWED_USERS=${DISINTO_CHAT_ALLOWED_USERS:-}
+      # Cost caps / rate limiting (#711)
+      - CHAT_MAX_REQUESTS_PER_HOUR=${CHAT_MAX_REQUESTS_PER_HOUR:-60}
+      - CHAT_MAX_REQUESTS_PER_DAY=${CHAT_MAX_REQUESTS_PER_DAY:-500}
+      - CHAT_MAX_TOKENS_PER_DAY=${CHAT_MAX_TOKENS_PER_DAY:-1000000}
     volumes:
       - ./docker/Caddyfile:/etc/caddy/Caddyfile
       - caddy_data:/data
@@ -622,6 +631,8 @@ COMPOSEEOF
       - ./secrets/tunnel_key:/run/secrets/tunnel_key:ro
       - ${CLAUDE_SHARED_DIR:-/var/lib/disinto/claude-shared}:${CLAUDE_SHARED_DIR:-/var/lib/disinto/claude-shared}
       - ${CLAUDE_CONFIG_FILE:-${HOME}/.claude.json}:/home/agent/.claude.json:ro
+      # Chat history persistence (merged from chat container, #1083)
+      - ${CHAT_HISTORY_DIR:-./state/chat-history}:/var/lib/chat/history
     healthcheck:
       test: ["CMD", "curl", "-fsS", "http://localhost:2019/config/"]
       interval: 30s
@@ -670,64 +681,12 @@ COMPOSEEOF
       - disinto-net
     command: ["echo", "staging slot — replace with project image"]
 
-  # Chat container — Claude chat UI backend (#705)
-  # Internal service only; edge proxy routes to chat:8080
-  # Sandbox hardened per #706 — no docker.sock, read-only rootfs, minimal caps
-  chat:
-    build:
-      context: ./docker/chat
-      dockerfile: Dockerfile
-    container_name: disinto-chat
-    restart: unless-stopped
-    read_only: true
-    tmpfs:
-      - /tmp:size=64m
-    security_opt:
-      - no-new-privileges:true
-    cap_drop:
-      - ALL
-    pids_limit: 128
-    mem_limit: 512m
-    memswap_limit: 512m
-    volumes:
-      # Mount claude binary from host (same as agents)
-      - ${CLAUDE_BIN_DIR}:/usr/local/bin/claude:ro
-      # Throwaway named volume for chat config (isolated from host ~/.claude)
-      - chat-config:/var/chat/config
-      # Chat history persistence: per-user NDJSON files on bind-mounted host volume
-      - ${CHAT_HISTORY_DIR:-./state/chat-history}:/var/lib/chat/history
-    environment:
-      CHAT_HOST: "0.0.0.0"
-      CHAT_PORT: "8080"
-      FORGE_URL: http://forgejo:3000
-      CHAT_OAUTH_CLIENT_ID: ${CHAT_OAUTH_CLIENT_ID:-}
-      CHAT_OAUTH_CLIENT_SECRET: ${CHAT_OAUTH_CLIENT_SECRET:-}
-      EDGE_TUNNEL_FQDN: ${EDGE_TUNNEL_FQDN:-}
-      EDGE_TUNNEL_FQDN_CHAT: ${EDGE_TUNNEL_FQDN_CHAT:-}
-      EDGE_ROUTING_MODE: ${EDGE_ROUTING_MODE:-subpath}
-      DISINTO_CHAT_ALLOWED_USERS: ${DISINTO_CHAT_ALLOWED_USERS:-}
-      # Shared secret for Caddy forward_auth verify endpoint (#709)
-      FORWARD_AUTH_SECRET: ${FORWARD_AUTH_SECRET:-}
-      # Cost caps / rate limiting (#711)
-      CHAT_MAX_REQUESTS_PER_HOUR: ${CHAT_MAX_REQUESTS_PER_HOUR:-60}
-      CHAT_MAX_REQUESTS_PER_DAY: ${CHAT_MAX_REQUESTS_PER_DAY:-500}
-      CHAT_MAX_TOKENS_PER_DAY: ${CHAT_MAX_TOKENS_PER_DAY:-1000000}
-    healthcheck:
-      test: ["CMD", "python3", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8080/health')"]
-      interval: 30s
-      timeout: 5s
-      retries: 3
-      start_period: 10s
-    networks:
-      - disinto-net
-
 volumes:
   forgejo-data:
   woodpecker-data:
   agent-data:
   project-repos:
   caddy_data:
-  chat-config:
 
 networks:
   disinto-net:
@@ -786,7 +745,7 @@ COMPOSEEOF
   # In build mode, replace image: with build: for locally-built images
   if [ "$use_build" = true ]; then
     sed -i '/^    image: ghcr\.io\/disinto\/agents:/{s|image: ghcr\.io/disinto/agents:.*|build:\n      context: .\n      dockerfile: docker/agents/Dockerfile\n    pull_policy: build|}' "$compose_file"
-    sed -i '/^    image: ghcr\.io\/disinto\/edge:/{s|image: ghcr\.io/disinto/edge:.*|build: ./docker/edge\n    pull_policy: build|}' "$compose_file"
+    sed -i '/^    image: ghcr\.io\/disinto\/edge:/{s|image: ghcr\.io/disinto/edge:.*|build:\n      context: .\n      dockerfile: docker/edge/Dockerfile\n    pull_policy: build|}' "$compose_file"
   fi
 
   echo "Created: ${compose_file}"
@@ -864,22 +823,22 @@ _generate_caddyfile_subpath() {
         reverse_proxy staging:80
     }
 
-    # Chat service — reverse proxy to disinto-chat backend (#705)
+    # Chat service — reverse proxy to in-process chat server (#705, #1083)
     # OAuth routes bypass forward_auth — unauthenticated users need these (#709)
     handle /chat/login {
-        reverse_proxy chat:8080
+        reverse_proxy 127.0.0.1:8080
     }
     handle /chat/oauth/callback {
-        reverse_proxy chat:8080
+        reverse_proxy 127.0.0.1:8080
     }
     # Defense-in-depth: forward_auth stamps X-Forwarded-User from session (#709)
     handle /chat/* {
-        forward_auth chat:8080 {
+        forward_auth 127.0.0.1:8080 {
             uri /chat/auth/verify
             copy_headers X-Forwarded-User
             header_up X-Forward-Auth-Secret {$FORWARD_AUTH_SECRET}
         }
-        reverse_proxy chat:8080
+        reverse_proxy 127.0.0.1:8080
     }
 }
 CADDYFILEEOF
@@ -912,18 +871,18 @@ _generate_caddyfile_subdomain() {
 # Chat — with forward_auth (#709, on its own host)
 {$EDGE_TUNNEL_FQDN_CHAT} {
     handle /login {
-        reverse_proxy chat:8080
+        reverse_proxy 127.0.0.1:8080
     }
     handle /oauth/callback {
-        reverse_proxy chat:8080
+        reverse_proxy 127.0.0.1:8080
     }
     handle /* {
-        forward_auth chat:8080 {
+        forward_auth 127.0.0.1:8080 {
             uri /auth/verify
             copy_headers X-Forwarded-User
             header_up X-Forward-Auth-Secret {$FORWARD_AUTH_SECRET}
         }
-        reverse_proxy chat:8080
+        reverse_proxy 127.0.0.1:8080
     }
 }
 CADDYFILEEOF