fireclaw 0.1.2 → 0.2.0

package/bin/vm-setup

@@ -8,13 +8,15 @@ CMD_NAME="${VM_SETUP_CMD_NAME:-fireclaw setup}"
 
 usage() {
   cat <<EOF
-Usage: $CMD_NAME --instance <id> --telegram-token <token> [options]
+Usage: $CMD_NAME --instance <id> [options]
 
 Options:
-  --telegram-users <csv>
-  --model <id>                         (default: anthropic/claude-opus-4-6)
+  --telegram-token <token>             (optional; omit for a local-only gateway with Telegram disabled)
+  --telegram-users <csv>               (required with --telegram-token)
+  --model <id>                         (default: openai/gpt-5.5)
   --skills <csv>                       (default: github,tmux,coding-agent,session-logs,skill-creator)
   --openclaw-image <image>
+  --host-port <n>                      (default: first free port above BASE_PORT)
   --vm-vcpu <n>                        (default: 4)
   --vm-mem-mib <n>                     (default: 8192)
   --disk-size <size>                   (default: 40G)
@@ -30,16 +32,48 @@ Options:
 EOF
 }
 
+require_option_value() {
+  local opt="$1"
+  local value="${2-}"
+  [[ -n "$value" && "$value" != --* ]] || die "Missing value for $opt"
+}
+
+validate_no_newline() {
+  local name="$1"
+  local value="$2"
+  [[ "$value" != *$'\n'* && "$value" != *$'\r'* ]] || die "$name must not contain newlines"
+}
+
+validate_positive_int() {
+  local name="$1"
+  local value="$2"
+  [[ "$value" =~ ^[0-9]+$ ]] || die "$name must be numeric"
+  (( 10#$value > 0 )) || die "$name must be greater than zero"
+}
+
+csv_values() {
+  printf '%s' "$1" | tr ',' '\n' | sed 's/^[[:space:]]*//; s/[[:space:]]*$//; /^$/d'
+}
+
+write_kv() {
+  local file="$1"
+  local key="$2"
+  local value="$3"
+  validate_no_newline "$key" "$value"
+  printf "%s=%s\n" "$key" "$value" >> "$file"
+}
+
 INSTANCE=""
 TELEGRAM_TOKEN=""
 TELEGRAM_USERS=""
-MODEL="${MODEL:-anthropic/claude-opus-4-6}"
+MODEL="${MODEL:-openai/gpt-5.5}"
 SKILLS="${SKILLS:-github,tmux,coding-agent,session-logs,skill-creator}"
 OPENCLAW_IMAGE="$OPENCLAW_IMAGE_DEFAULT"
 VM_VCPU="${VM_VCPU:-4}"
 VM_MEM_MIB="${VM_MEM_MIB:-8192}"
 DISK_SIZE="${DISK_SIZE:-40G}"
 API_SOCK="${API_SOCK:-}"
+HOST_PORT_OVERRIDE=""
 BASE_IMAGES_DIR="${BASE_IMAGES_DIR:-/srv/firecracker/base/images}"
 BASE_KERNEL="${BASE_KERNEL:-$BASE_IMAGES_DIR/vmlinux}"
 BASE_ROOTFS="${BASE_ROOTFS:-$BASE_IMAGES_DIR/rootfs.ext4}"
@@ -49,25 +83,63 @@ SKIP_BROWSER_INSTALL="false"
 ANTHROPIC_API_KEY="${ANTHROPIC_API_KEY:-}"
 OPENAI_API_KEY="${OPENAI_API_KEY:-}"
 MINIMAX_API_KEY="${MINIMAX_API_KEY:-}"
+SETUP_COMPLETE="false"
+CLEANUP_ARMED="false"
+
+cleanup_failed_setup() {
+  local rc="$1"
+  [[ "$rc" -eq 0 || "$SETUP_COMPLETE" == "true" ]] && return 0
+  [[ "$CLEANUP_ARMED" == "true" ]] || return 0
+  set +e
+  if [[ -n "${INSTANCE:-}" && "$INSTANCE" =~ ^[a-z0-9_-]+$ ]]; then
+    warn "Setup failed; cleaning up partial instance: $INSTANCE"
+    local vm_svc proxy_svc
+    vm_svc="$(vm_service "$INSTANCE")"
+    proxy_svc="$(proxy_service "$INSTANCE")"
+    systemctl stop "$proxy_svc" 2>/dev/null || true
+    systemctl stop "$vm_svc" 2>/dev/null || true
+    systemctl disable "$proxy_svc" 2>/dev/null || true
+    systemctl disable "$vm_svc" 2>/dev/null || true
+    rm -f "/etc/systemd/system/$proxy_svc" "/etc/systemd/system/$vm_svc"
+    systemctl daemon-reload 2>/dev/null || true
+    systemctl reset-failed "$proxy_svc" "$vm_svc" 2>/dev/null || true
+    if [[ -n "${VM_TAP:-}" ]]; then
+      ip link set "$VM_TAP" down 2>/dev/null || true
+      ip link del "$VM_TAP" 2>/dev/null || true
+    fi
+    if [[ -n "${API_SOCK:-}" ]]; then
+      rm -f "$API_SOCK"
+    fi
+    if [[ -n "${inst_dir:-}" ]]; then
+      rm -rf "$inst_dir"
+    fi
+    if [[ -n "${fc_dir:-}" ]]; then
+      rm -rf "$fc_dir"
+    fi
+  fi
+}
+
+trap 'rc=$?; cleanup_failed_setup "$rc"; exit "$rc"' EXIT
 
 while [[ $# -gt 0 ]]; do
   case "$1" in
-    --instance) INSTANCE="$2"; shift 2 ;;
-    --telegram-token) TELEGRAM_TOKEN="$2"; shift 2 ;;
-    --telegram-users) TELEGRAM_USERS="$2"; shift 2 ;;
-    --model) MODEL="$2"; shift 2 ;;
-    --skills) SKILLS="$2"; shift 2 ;;
-    --openclaw-image) OPENCLAW_IMAGE="$2"; shift 2 ;;
-    --vm-vcpu) VM_VCPU="$2"; shift 2 ;;
-    --vm-mem-mib) VM_MEM_MIB="$2"; shift 2 ;;
-    --disk-size) DISK_SIZE="$2"; shift 2 ;;
-    --api-sock) API_SOCK="$2"; shift 2 ;;
-    --base-kernel) BASE_KERNEL="$2"; shift 2 ;;
-    --base-rootfs) BASE_ROOTFS="$2"; shift 2 ;;
-    --base-initrd) BASE_INITRD="$2"; shift 2 ;;
-    --anthropic-api-key) ANTHROPIC_API_KEY="$2"; shift 2 ;;
-    --openai-api-key) OPENAI_API_KEY="$2"; shift 2 ;;
-    --minimax-api-key) MINIMAX_API_KEY="$2"; shift 2 ;;
+    --instance) require_option_value "$1" "${2-}"; INSTANCE="$2"; shift 2 ;;
+    --telegram-token) require_option_value "$1" "${2-}"; TELEGRAM_TOKEN="$2"; shift 2 ;;
+    --telegram-users) require_option_value "$1" "${2-}"; TELEGRAM_USERS="$2"; shift 2 ;;
+    --model) require_option_value "$1" "${2-}"; MODEL="$2"; shift 2 ;;
+    --skills) require_option_value "$1" "${2-}"; SKILLS="$2"; shift 2 ;;
+    --openclaw-image) require_option_value "$1" "${2-}"; OPENCLAW_IMAGE="$2"; shift 2 ;;
+    --host-port) require_option_value "$1" "${2-}"; HOST_PORT_OVERRIDE="$2"; shift 2 ;;
+    --vm-vcpu) require_option_value "$1" "${2-}"; VM_VCPU="$2"; shift 2 ;;
+    --vm-mem-mib) require_option_value "$1" "${2-}"; VM_MEM_MIB="$2"; shift 2 ;;
+    --disk-size) require_option_value "$1" "${2-}"; DISK_SIZE="$2"; shift 2 ;;
+    --api-sock) require_option_value "$1" "${2-}"; API_SOCK="$2"; shift 2 ;;
+    --base-kernel) require_option_value "$1" "${2-}"; BASE_KERNEL="$2"; shift 2 ;;
+    --base-rootfs) require_option_value "$1" "${2-}"; BASE_ROOTFS="$2"; shift 2 ;;
+    --base-initrd) require_option_value "$1" "${2-}"; BASE_INITRD="$2"; shift 2 ;;
+    --anthropic-api-key) require_option_value "$1" "${2-}"; ANTHROPIC_API_KEY="$2"; shift 2 ;;
+    --openai-api-key) require_option_value "$1" "${2-}"; OPENAI_API_KEY="$2"; shift 2 ;;
+    --minimax-api-key) require_option_value "$1" "${2-}"; MINIMAX_API_KEY="$2"; shift 2 ;;
     --skip-browser-install) SKIP_BROWSER_INSTALL="true"; shift ;;
     -h|--help) usage; exit 0 ;;
     *) die "Unknown option: $1" ;;
@@ -75,13 +147,47 @@ while [[ $# -gt 0 ]]; do
 done
 
 [[ -n "$INSTANCE" ]] || die "Missing --instance"
-[[ -n "$TELEGRAM_TOKEN" ]] || die "Missing --telegram-token"
+if [[ -n "$TELEGRAM_TOKEN" ]]; then
+  [[ -n "$TELEGRAM_USERS" ]] || die "Missing --telegram-users (required with --telegram-token)"
+  [[ -n "$(csv_values "$TELEGRAM_USERS")" ]] || die "--telegram-users must include at least one allowed Telegram user ID"
+elif [[ -n "$TELEGRAM_USERS" ]]; then
+  die "--telegram-users requires --telegram-token"
+fi
 validate_instance_id "$INSTANCE"
+validate_positive_int "VM_VCPU" "$VM_VCPU"
+validate_positive_int "VM_MEM_MIB" "$VM_MEM_MIB"
+validate_no_newline "TELEGRAM_TOKEN" "$TELEGRAM_TOKEN"
+validate_no_newline "TELEGRAM_USERS" "$TELEGRAM_USERS"
+validate_no_newline "MODEL" "$MODEL"
+validate_no_newline "SKILLS" "$SKILLS"
+validate_no_newline "OPENCLAW_IMAGE" "$OPENCLAW_IMAGE"
+validate_no_newline "DISK_SIZE" "$DISK_SIZE"
+validate_no_newline "API_SOCK" "$API_SOCK"
+validate_no_newline "BASE_KERNEL" "$BASE_KERNEL"
+validate_no_newline "BASE_ROOTFS" "$BASE_ROOTFS"
+validate_no_newline "BASE_INITRD" "$BASE_INITRD"
+validate_no_newline "ANTHROPIC_API_KEY" "$ANTHROPIC_API_KEY"
+validate_no_newline "OPENAI_API_KEY" "$OPENAI_API_KEY"
+validate_no_newline "MINIMAX_API_KEY" "$MINIMAX_API_KEY"
+if [[ -n "$HOST_PORT_OVERRIDE" ]]; then
+  validate_no_newline "HOST_PORT" "$HOST_PORT_OVERRIDE"
+fi
+VM_VCPU=$((10#$VM_VCPU))
+VM_MEM_MIB=$((10#$VM_MEM_MIB))
+require_model_provider_key "$MODEL"
 
 require_root
-for c in firecracker systemctl ip iptables openssl jq cloud-localds ssh scp socat curl qemu-img; do
+for c in firecracker systemctl ip bridge iptables openssl jq cloud-localds ssh scp socat curl qemu-img install flock; do
   require_cmd "$c"
 done
+FIRECRACKER_BIN="$(command -v firecracker)"
+[[ "$FIRECRACKER_BIN" = /* ]] || die "firecracker must resolve to an absolute executable path"
+[[ -e /dev/kvm && -r /dev/kvm && -w /dev/kvm ]] || die "/dev/kvm is not available (KVM is required to run Firecracker VMs)"
+
+mem_avail_mib="$(awk '/^MemAvailable:/ {print int($2/1024)}' /proc/meminfo 2>/dev/null || echo 0)"
+if (( mem_avail_mib > 0 && VM_MEM_MIB > mem_avail_mib )); then
+  warn "Requested VM memory (${VM_MEM_MIB} MiB) exceeds host MemAvailable (${mem_avail_mib} MiB); the VM may OOM the host"
+fi
 
 [[ -f "$BASE_KERNEL" ]] || die "Kernel not found: $BASE_KERNEL"
 [[ -f "$BASE_ROOTFS" ]] || die "Rootfs not found: $BASE_ROOTFS"
@@ -99,11 +205,30 @@ fc_dir="$(fc_instance_dir "$INSTANCE")"
 if [[ -z "$API_SOCK" ]]; then
   API_SOCK="$fc_dir/firecracker.socket"
 fi
+VM_SVC="$(vm_service "$INSTANCE")"
+PROXY_SVC="$(proxy_service "$INSTANCE")"
 [[ "$API_SOCK" = /* ]] || die "--api-sock must be an absolute path"
 [[ ! -e "$inst_dir" ]] || die "Instance already exists: $inst_dir"
-
-HOST_PORT="$(next_port)"
+[[ ! -e "$fc_dir" ]] || die "Firecracker assets already exist: $fc_dir"
+[[ ! -e "$API_SOCK" ]] || die "Firecracker API socket already exists: $API_SOCK"
+[[ ! -e "/etc/systemd/system/$VM_SVC" ]] || die "VM unit already exists: /etc/systemd/system/$VM_SVC"
+[[ ! -e "/etc/systemd/system/$PROXY_SVC" ]] || die "Proxy unit already exists: /etc/systemd/system/$PROXY_SVC"
+CLEANUP_ARMED="true"
+
+# Hold the allocation lock until the reservation is visible in the .env file,
+# so concurrent setups cannot pick the same port/IP.
+exec {ALLOC_LOCK_FD}>"$STATE_ROOT/.alloc.lock"
+flock "$ALLOC_LOCK_FD"
+
+if [[ -n "$HOST_PORT_OVERRIDE" ]]; then
+  ensure_host_port_available "$HOST_PORT_OVERRIDE"
+  HOST_PORT="$((10#$HOST_PORT_OVERRIDE))"
+else
+  HOST_PORT="$(next_port)"
+fi
 VM_IP="$(next_ip)"
+SUBNET_MASK_BITS="$(subnet_mask_bits)"
+BRIDGE_GATEWAY_IP="$(bridge_gateway_ip)"
 VM_OCTET="${VM_IP##*.}"
 SHORT_ID="$(echo "$INSTANCE" | tr -cd 'a-z0-9' | cut -c1-6)"
 [[ -n "$SHORT_ID" ]] || SHORT_ID="vm"
@@ -111,40 +236,50 @@ VM_TAP="t${SHORT_ID}${VM_OCTET}"
 VM_MAC="$(printf '06:fc:00:10:00:%02x' "$VM_OCTET")"
 GATEWAY_TOKEN="$(openssl rand -hex 32)"
 
+if ip link show "$VM_TAP" >/dev/null 2>&1; then
+  die "TAP device already exists: $VM_TAP (leftover from a previous instance? remove it with: sudo ip link del $VM_TAP)"
+fi
+
 mkdir -p "$inst_dir/config" "$inst_dir/workspace"
+chmod 700 "$inst_dir" "$inst_dir/config" "$inst_dir/workspace"
 mkdir -p "$fc_dir/images" "$fc_dir/config" "$fc_dir/logs"
+chmod 700 "$fc_dir" "$fc_dir/images" "$fc_dir/config" "$fc_dir/logs"
 mkdir -p "$(dirname "$API_SOCK")"
 
-cat > "$inst_dir/.env" <<EOF
-INSTANCE_ID=$INSTANCE
-HOST_PORT=$HOST_PORT
-VM_IP=$VM_IP
-VM_TAP=$VM_TAP
-VM_MAC=$VM_MAC
-GATEWAY_TOKEN=$GATEWAY_TOKEN
-MODEL=$MODEL
-SKILLS=$SKILLS
-TELEGRAM_USERS=$TELEGRAM_USERS
-OPENCLAW_IMAGE=$OPENCLAW_IMAGE
-VM_VCPU=$VM_VCPU
-VM_MEM_MIB=$VM_MEM_MIB
-DISK_SIZE=$DISK_SIZE
-API_SOCK=$API_SOCK
-SSH_KEY_PATH=$SSH_KEY_PATH
-SKIP_BROWSER_INSTALL=$SKIP_BROWSER_INSTALL
-ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY
-OPENAI_API_KEY=$OPENAI_API_KEY
-MINIMAX_API_KEY=$MINIMAX_API_KEY
-EOF
-
-echo "$GATEWAY_TOKEN" > "$inst_dir/.token"
-chmod 600 "$inst_dir/.token"
+ENV_FILE="$(instance_env "$INSTANCE")"
+install -m 600 /dev/null "$ENV_FILE"
+write_kv "$ENV_FILE" FIRECLAW_STATE_FORMAT "plain-v1"
+write_kv "$ENV_FILE" INSTANCE_ID "$INSTANCE"
+write_kv "$ENV_FILE" HOST_PORT "$HOST_PORT"
+write_kv "$ENV_FILE" VM_IP "$VM_IP"
+write_kv "$ENV_FILE" VM_TAP "$VM_TAP"
+write_kv "$ENV_FILE" VM_MAC "$VM_MAC"
+write_kv "$ENV_FILE" GATEWAY_TOKEN "$GATEWAY_TOKEN"
+write_kv "$ENV_FILE" MODEL "$MODEL"
+write_kv "$ENV_FILE" SKILLS "$SKILLS"
+write_kv "$ENV_FILE" TELEGRAM_USERS "$TELEGRAM_USERS"
+write_kv "$ENV_FILE" OPENCLAW_IMAGE "$OPENCLAW_IMAGE"
+write_kv "$ENV_FILE" VM_VCPU "$VM_VCPU"
+write_kv "$ENV_FILE" VM_MEM_MIB "$VM_MEM_MIB"
+write_kv "$ENV_FILE" DISK_SIZE "$DISK_SIZE"
+write_kv "$ENV_FILE" API_SOCK "$API_SOCK"
+write_kv "$ENV_FILE" SSH_KEY_PATH "$SSH_KEY_PATH"
+write_kv "$ENV_FILE" SKIP_BROWSER_INSTALL "$SKIP_BROWSER_INSTALL"
+write_kv "$ENV_FILE" ANTHROPIC_API_KEY "$ANTHROPIC_API_KEY"
+write_kv "$ENV_FILE" OPENAI_API_KEY "$OPENAI_API_KEY"
+write_kv "$ENV_FILE" MINIMAX_API_KEY "$MINIMAX_API_KEY"
+
+exec {ALLOC_LOCK_FD}>&-
+
+install -m 600 /dev/null "$inst_dir/.token"
+printf '%s\n' "$GATEWAY_TOKEN" > "$inst_dir/.token"
 
 log "Preparing VM disk and assets"
 cp --reflink=auto "$BASE_ROOTFS" "$fc_dir/images/rootfs.ext4" 2>/dev/null || cp "$BASE_ROOTFS" "$fc_dir/images/rootfs.ext4"
+chmod 600 "$fc_dir/images/rootfs.ext4"
 if [[ -n "$DISK_SIZE" ]]; then
   log "Resizing rootfs to $DISK_SIZE"
-  qemu-img resize "$fc_dir/images/rootfs.ext4" "$DISK_SIZE" >/dev/null
+  qemu-img resize -f raw "$fc_dir/images/rootfs.ext4" "$DISK_SIZE" >/dev/null
 fi
 cp "$BASE_KERNEL" "$fc_dir/images/vmlinux"
 
@@ -168,12 +303,8 @@ users:
       - $SSH_PUB_KEY
 package_update: true
 packages:
-  - docker.io
   - jq
   - curl
-runcmd:
-  - [ systemctl, enable, docker ]
-  - [ systemctl, start, docker ]
 EOF
 
 cat > "$fc_dir/config/meta-data" <<EOF
@@ -190,10 +321,10 @@ ethernets:
     set-name: eth0
     dhcp4: false
     addresses:
-      - $VM_IP/24
+      - $VM_IP/$SUBNET_MASK_BITS
     routes:
       - to: 0.0.0.0/0
-        via: 172.16.0.1
+        via: $BRIDGE_GATEWAY_IP
     nameservers:
       addresses: [1.1.1.1,8.8.8.8]
 EOF
@@ -237,21 +368,29 @@ SUBNET_CIDR="$SUBNET_CIDR"
 VM_TAP="$VM_TAP"
 API_SOCK="$API_SOCK"
 CONFIG_JSON="$fc_dir/config/vm-config.json"
+FIRECRACKER_BIN="$FIRECRACKER_BIN"
 
 if ! ip link show "\$BRIDGE_NAME" >/dev/null 2>&1; then
   ip link add "\$BRIDGE_NAME" type bridge
 fi
-ip addr add "\$BRIDGE_ADDR" dev "\$BRIDGE_NAME" 2>/dev/null || true
+ip -4 addr show dev "\$BRIDGE_NAME" | grep -Fq " \$BRIDGE_ADDR " || ip addr add "\$BRIDGE_ADDR" dev "\$BRIDGE_NAME"
 ip link set "\$BRIDGE_NAME" up
 sysctl -w net.ipv4.ip_forward=1 >/dev/null
 iptables -t nat -C POSTROUTING -s "\$SUBNET_CIDR" ! -o "\$BRIDGE_NAME" -j MASQUERADE 2>/dev/null || iptables -t nat -A POSTROUTING -s "\$SUBNET_CIDR" ! -o "\$BRIDGE_NAME" -j MASQUERADE
+# Defense-in-depth only: matches solely where br_netfilter is enabled.
+iptables -C FORWARD -i "\$BRIDGE_NAME" -o "\$BRIDGE_NAME" -j DROP 2>/dev/null || iptables -I FORWARD 1 -i "\$BRIDGE_NAME" -o "\$BRIDGE_NAME" -j DROP
 
-ip tuntap add dev "\$VM_TAP" mode tap user root 2>/dev/null || true
+if ! ip link show "\$VM_TAP" >/dev/null 2>&1; then
+  ip tuntap add dev "\$VM_TAP" mode tap user root
+fi
 ip link set "\$VM_TAP" master "\$BRIDGE_NAME"
+# Isolated bridge ports cannot exchange frames (IP or ARP) with each other,
+# only with the bridge/host itself — this is the actual VM-to-VM isolation.
+bridge link set dev "\$VM_TAP" isolated on
 ip link set "\$VM_TAP" up
 
 rm -f "\$API_SOCK"
-exec /usr/local/bin/firecracker --api-sock "\$API_SOCK" --config-file "\$CONFIG_JSON"
+exec "\$FIRECRACKER_BIN" --api-sock "\$API_SOCK" --config-file "\$CONFIG_JSON"
 EOF
 chmod +x "$fc_dir/start-vm.sh"
 
@@ -259,14 +398,27 @@ cat > "$fc_dir/stop-vm.sh" <<EOF
 #!/usr/bin/env bash
 set -euo pipefail
 VM_TAP="$VM_TAP"
+API_SOCK="$API_SOCK"
+
+# Ask the guest to shut down cleanly (systemd maps Ctrl-Alt-Del to a clean
+# reboot, which exits Firecracker) before systemd kills the VMM, so the
+# rootfs ext4 is not torn down mid-write.
+if [[ -S "\$API_SOCK" ]]; then
+  if curl -fsS --max-time 2 --unix-socket "\$API_SOCK" -X PUT http://localhost/actions \\
+      -H 'Content-Type: application/json' -d '{"action_type": "SendCtrlAltDel"}' >/dev/null 2>&1; then
+    for _ in \$(seq 1 20); do
+      curl -fsS --max-time 1 --unix-socket "\$API_SOCK" http://localhost/ >/dev/null 2>&1 || break
+      sleep 1
+    done
+  fi
+fi
+
 ip link set "\$VM_TAP" down 2>/dev/null || true
 ip link del "\$VM_TAP" 2>/dev/null || true
+rm -f "\$API_SOCK"
 EOF
 chmod +x "$fc_dir/stop-vm.sh"
 
-VM_SVC="$(vm_service "$INSTANCE")"
-PROXY_SVC="$(proxy_service "$INSTANCE")"
-
 cat > "/etc/systemd/system/$VM_SVC" <<EOF
 [Unit]
 Description=Firecracker VM ($INSTANCE)
@@ -277,7 +429,8 @@ Wants=network-online.target
 Type=simple
 ExecStart=$fc_dir/start-vm.sh
 ExecStop=$fc_dir/stop-vm.sh
-Restart=on-failure
+TimeoutStopSec=45
+Restart=always
 RestartSec=2
 
 [Install]
@@ -289,29 +442,33 @@ cat > "/etc/systemd/system/$PROXY_SVC" <<EOF
 Description=Localhost proxy for VM ($INSTANCE)
 After=$VM_SVC
 Requires=$VM_SVC
+PartOf=$VM_SVC
 
 [Service]
 Type=simple
+DynamicUser=yes
 ExecStart=/usr/bin/socat TCP-LISTEN:$HOST_PORT,bind=127.0.0.1,reuseaddr,fork TCP:$VM_IP:18789
+SuccessExitStatus=143
 Restart=always
 RestartSec=2
 
 [Install]
-WantedBy=multi-user.target
+WantedBy=multi-user.target $VM_SVC
 EOF
 
 systemctl daemon-reload
 systemctl enable --now "$VM_SVC"
 
 log "Waiting for SSH on $VM_IP"
-if ! wait_for_ssh "$VM_IP" "$SSH_KEY_PATH" 180; then
+if ! wait_for_ssh "$VM_IP" "$SSH_KEY_PATH" 180 "$INSTANCE"; then
   die "VM booted but SSH was not reachable"
 fi
 
 PROVISION_VARS="$inst_dir/provision.vars"
-: > "$PROVISION_VARS"
-kv() { printf "%s=%q\n" "$1" "$2" >> "$PROVISION_VARS"; }
+install -m 600 /dev/null "$PROVISION_VARS"
+kv() { write_kv "$PROVISION_VARS" "$1" "$2"; }
 
+kv FIRECLAW_STATE_FORMAT "plain-v1"
 kv INSTANCE_ID "$INSTANCE"
 kv TELEGRAM_TOKEN "$TELEGRAM_TOKEN"
 kv TELEGRAM_USERS "$TELEGRAM_USERS"
@@ -327,44 +484,30 @@ kv MINIMAX_API_KEY "$MINIMAX_API_KEY"
 
 [[ -f "$REPO_ROOT/scripts/provision-guest.sh" ]] || die "Missing: $REPO_ROOT/scripts/provision-guest.sh"
 
-scp -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=accept-new -o UserKnownHostsFile=/dev/null \
+scp -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=accept-new -o UserKnownHostsFile="$(ssh_known_hosts_file "$INSTANCE")" \
   "$REPO_ROOT/scripts/provision-guest.sh" "ubuntu@$VM_IP:/tmp/provision-guest.sh"
-scp -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=accept-new -o UserKnownHostsFile=/dev/null \
+scp -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=accept-new -o UserKnownHostsFile="$(ssh_known_hosts_file "$INSTANCE")" \
   "$PROVISION_VARS" "ubuntu@$VM_IP:/tmp/provision.vars"
 
-ssh -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=accept-new -o UserKnownHostsFile=/dev/null \
+ssh -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=accept-new -o UserKnownHostsFile="$(ssh_known_hosts_file "$INSTANCE")" \
   "ubuntu@$VM_IP" "sudo bash /tmp/provision-guest.sh /tmp/provision.vars"
 
+# Guest provisioning succeeded: from here on a failure must not destroy the
+# instance — it is recoverable with `fireclaw provision` / `fireclaw start`.
+CLEANUP_ARMED="false"
+
 systemctl enable --now "$PROXY_SVC"
 
-host_health_ok="false"
-guest_health_ok="false"
-GUEST_HEALTH_SCRIPT="$(guest_health_script "$INSTANCE")"
-for _ in {1..30}; do
-  if [[ "$guest_health_ok" != "true" ]] && check_guest_health "$INSTANCE" "$VM_IP" "$SSH_KEY_PATH"; then
-    guest_health_ok="true"
-  fi
-  if [[ "$host_health_ok" != "true" ]] && curl -fsS "http://127.0.0.1:$HOST_PORT/health" >/dev/null 2>&1; then
-    host_health_ok="true"
-  fi
-  if [[ "$guest_health_ok" == "true" && "$host_health_ok" == "true" ]]; then
-    break
-  fi
-  sleep 2
-done
+if ! wait_for_instance_health "$INSTANCE" "$VM_IP" "$HOST_PORT" "$SSH_KEY_PATH" 30; then
+  warn "The instance was kept for inspection: sudo fireclaw status $INSTANCE"
+  die "Health checks did not pass for $INSTANCE after provisioning (retry with: sudo fireclaw provision $INSTANCE)"
+fi
+SETUP_COMPLETE="true"
 
 echo "✓ VM instance configured"
 echo "  Instance: $INSTANCE"
 echo "  VM IP:    $VM_IP"
 echo "  Port:     $HOST_PORT"
-echo "  Token:    $GATEWAY_TOKEN"
-if [[ "$guest_health_ok" == "true" && "$host_health_ok" == "true" ]]; then
-  echo "  Health:   up (guest + proxy)"
-elif [[ "$guest_health_ok" == "true" ]]; then
-  echo "  Health:   guest up, proxy pending"
-elif [[ "$host_health_ok" == "true" ]]; then
-  echo "  Health:   proxy up (guest check pending)"
-else
-  echo "  Health:   pending (guest script: $GUEST_HEALTH_SCRIPT)"
-fi
+echo "  Token:    (print with: sudo fireclaw token $INSTANCE)"
+echo "  Health:   up (guest + proxy)"
 echo "  Status:   $(systemctl is-active "$VM_SVC") / $(systemctl is-active "$PROXY_SVC")"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "fireclaw",
-  "version": "0.1.2",
+  "version": "0.2.0",
   "description": "Firecracker microVM control plane for isolated OpenClaw instances",
   "bin": {
     "fireclaw": "bin/fireclaw"
@@ -11,8 +11,9 @@
     "README.md"
   ],
   "scripts": {
-    "lint:shell": "bash -n bin/fireclaw && bash -n bin/vm-common.sh && bash -n bin/vm-setup && bash -n bin/vm-ctl && bash -n scripts/provision-guest.sh",
-    "test": "npm run lint:shell",
+    "lint:shell": "bash -n bin/fireclaw && bash -n bin/vm-common.sh && bash -n bin/vm-setup && bash -n bin/vm-ctl && bash -n bin/vm-provision && bash -n scripts/provision-guest.sh && bash -n tests/unit/vm-common.test.sh",
+    "test:unit": "bash tests/unit/vm-common.test.sh",
+    "test": "npm run lint:shell && npm run test:unit",
     "pack:check": "npm pack --dry-run",
     "prepublishOnly": "npm test && npm run pack:check"
   },
@@ -31,7 +32,7 @@
   "bugs": {
     "url": "https://github.com/bchewy/fireclaw/issues"
   },
-  "license": "UNLICENSED",
+  "license": "MIT",
   "author": "bchewy",
   "engines": {
     "node": ">=18"