install.sh spin() has no trap handler — Ctrl+C orphans background processes and leaks temp files

[latenighthackathon]

Description

install.sh has zero trap statements in the entire 550+ line script. The spin() function (line 185) backgrounds a command and polls it with a spinner animation, but if the user presses Ctrl+C during execution:

1. The backgrounded process (npm install, docker build, git clone, etc.) continues running
2. The temp file created by mktemp on line 188 is never cleaned up
3. Any partially-written state (node_modules, Docker layers, cloned repos) is left behind

What happens: Ctrl+C during installation leaves orphaned processes running and temp files in /tmp.

What should happen: The script should register a cleanup trap that kills backgrounded processes and removes temp files on SIGINT/SIGTERM.

Reproduction Steps

1. Run the installer: curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash
2. During any spinner step (e.g., "Preparing OpenClaw package", "Installing dependencies"), press Ctrl+C
3. Check for orphaned processes:

   ps aux | grep -E "npm install|docker build|git clone"

4. Check for leaked temp files:

   ls /tmp/tmp.*

Environment

• Code review of main branch (commit HEAD as of 2026-03-26)
• Affected file: install.sh — entire file

Logs

No trap in the entire script:

$ grep -n 'trap ' install.sh
# (no output)

spin() function — backgrounds command, no cleanup (lines 185–212):

spin() {
  # ...
  local log
  log=$(mktemp)              # temp file created
  "$@" >"$log" 2>&1 &        # command backgrounded
  local pid=$! i=0

  while kill -0 "$pid" 2>/dev/null; do
    printf "\r  %s  %s" "${frames[$((i++ % 10))]}" "$msg"
    sleep 0.08
  done

  # ... status check ...
  rm -f "$log"               # only cleaned up on normal exit
  return $status
}

Expected pattern:

_cleanup_pids=()
_cleanup_files=()

cleanup() {
  for pid in "${_cleanup_pids[@]}"; do
    kill "$pid" 2>/dev/null
  done
  for f in "${_cleanup_files[@]}"; do
    rm -f "$f"
  done
}
trap cleanup EXIT SIGINT SIGTERM

spin() {
  local log
  log=$(mktemp)
  _cleanup_files+=("$log")
  "$@" >"$log" 2>&1 &
  local pid=$!
  _cleanup_pids+=("$pid")
  # ...
}

Checklist

• I confirmed this bug is reproducible
• I searched existing issues and this is not a duplicate