Skip to content

v-modal/vmodal_mcp / skills/vmodal-track

vmodal-track

Verified

Track a subject through VModal video collections: find every verified appearance of it and return a trail (where, when, for how long, in what order). Works for a description ("red taxis in Kyoto") or for one specific subject ("this taxi", "the same woman", "the one at 1:57"). Use when the user asks to track, follow, trace, or find everywhere / every time something appears, e.g. "Track red cars. Collection: city". For a single good match use vmodal-find instead. Requires the vmodal MCP (dev branch) and the vmodal-find skill.

FileKindSize
SKILL.mdskill md8.1 KB
reference.mdmarkdown3.0 KB
scripts/anchor_b64.shexecutablescript1.6 KB

scripts/anchor_b64.sh

View on GitHub
#!/usr/bin/env bash
help='
  Crop a subject out of a frame and print it as a small base64 JPEG, for
  search_video(image_query=...). The VModal API expects plain base64 image bytes
  (a URL is rejected with "cannot decode image_query base64").
  Output is ~3-4 KB of base64 (112 px wide, JPEG quality 40), small enough to pass in a tool call.
  Needs sips (macOS) or ImageMagick (magick/convert).

  anchor_b64.sh <frame.png> [x y w h]                  # default: img_b64
  anchor_b64.sh img_b64 /tmp/vmodal-find/city/a.png 225 190 210 160
  anchor_b64.sh img_b64 /tmp/vmodal-find/city/a.png     # whole frame
'

img_b64() {
  local help='
    ## Usage:
      anchor_b64.sh img_b64 /tmp/vmodal-find/city/a.png 225 190 210 160
  '
  local src="$1" x="${2:-}" y="${3:-}" w="${4:-}" h="${5:-}" tmp
  [[ -f "$src" ]] || { echo "missing file: $src" >&2; return 1; }
  tmp=$(mktemp -d)
  if command -v sips >/dev/null; then
    cp "$src" "$tmp/in.png"
    [[ -n "$h" ]] && sips -c "$h" "$w" --cropOffset "$y" "$x" "$tmp/in.png" --out "$tmp/in.png" >/dev/null
    sips -s format jpeg -s formatOptions 40 --resampleWidth 112 "$tmp/in.png" --out "$tmp/out.jpg" >/dev/null
  else
    local im; im=$(command -v magick || command -v convert) || { echo "need sips or ImageMagick" >&2; return 1; }
    local crop=(); [[ -n "$h" ]] && crop=(-crop "${w}x${h}+${x}+${y}" +repage)
    "$im" "$src" "${crop[@]}" -resize 112x -quality 40 "$tmp/out.jpg"
  fi
  base64 < "$tmp/out.jpg" | tr -d '\n'
  echo
  rm -rf "$tmp"
}

case "${1:-}" in
  ""|-h|--help|help) echo "$help" ;;
  img_b64) shift; img_b64 "$@" ;;
  *) img_b64 "$@" ;;
esac
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall