Ghost Decode

Ghost-Font Video Decoder

Ghost-font videos hide a message as a random-dot field: every frame is uniform noise, but the dots inside the letter shapes move against the background dots. This skill accumulates that motion into two images where the letters appear, then reads the message from them.

Hard rules — the whole job is ONE run producing TWO images

These rules exist because the #1 failure of this skill is over-processing: an agent that doesn't trust the output spawns a dozen diagnostic images, invents new algorithms, and hallucinates a message out of noise. Do not do that.

• Run the decoder exactly once. The algorithm below is correct and complete. Do not write a second decoder, try another method (temporal variance, phase correlation, sub-pixel warping, weighted accumulation, per-line crops…), or "improve" the pipeline.
• Produce exactly two images: revealed.png and revealed_heatmap.png. Create NO other images — no diagnostic maps, no crops, no re-thresholded or contrast-boosted variants. Extra images mean you are off the rails; stop.
• Never OCR a raw frame. Every single frame is pure noise; the message exists only in accumulated motion.
• Read the two images, then stop. Soft, rounded, blobby letters are the normal, correct output — not a reason to re-process. If you can read the word, report it.

Steps

1. Resolve the video path from $ARGUMENTS, the user's request, or the most recently modified video (.mp4, .mov, .avi, .webm) in the working directory. Ask only if several candidates are plausible.

2. Check dependencies: python -c "import cv2, numpy". If that fails, pip install -r "${CLAUDE_PLUGIN_ROOT}/requirements.txt" (or pip install opencv-python-headless numpy).

3. Decode — run once. Pick ONE:

   • If ${CLAUDE_PLUGIN_ROOT}/decode.py exists: python "${CLAUDE_PLUGIN_ROOT}/decode.py" "<video>" -o "<out-dir>"
   • Otherwise (plugin files not present in this environment): write the Decoder program at the bottom of this file verbatim to a scratch decode.py, then python decode.py "<video>" "<out-dir>".

   Either path writes exactly revealed.png and revealed_heatmap.png and nothing else. ${CLAUDE_PLUGIN_ROOT} is the plugin's install dir (on Windows PowerShell, $env:CLAUDE_PLUGIN_ROOT); if it expands empty, use the embedded Decoder instead.

4. Read the message. Read revealed.png with vision (it's a black background with the message in white); use revealed_heatmap.png to confirm a faint or merged glyph. The printed OCR hint line is only a rough hint from Tesseract — trust your own reading of the image over it. Mark any single ambiguous glyph (unclear: X).

Required response format

Show both images, then the text — nothing else:

![revealed.png](<absolute-path-to-revealed.png>)

![revealed_heatmap.png](<absolute-path-to-revealed_heatmap.png>)

Text in the video: **<RECOVERED TEXT>**

Use absolute local paths so the images render in chat. Do not claim success if the program did not run or you did not inspect revealed.png. If the mask genuinely has no letter shapes (just specks / a uniformly dark heatmap), say no text was recovered — still show the two images.

Troubleshooting (still one run, still two images)

• Weak or empty mask: rerun the SAME decoder once with --method farneback, and for high-fps clips add --stride 2. That is the only permitted retry. It still produces just the two images — do not switch algorithms or add diagnostic renders.
• Long video: add --max-frames 200; a few seconds of footage is enough.
• No Tesseract: fine — read the text from revealed.png yourself.

Decoder (write to a scratch decode.py only if the bundled one is absent)

import sys, os, shutil
import cv2, numpy as np

VIDEO = sys.argv[1] if len(sys.argv) > 1 else "video.mp4"
OUT = sys.argv[2] if len(sys.argv) > 2 else "out"
os.makedirs(OUT, exist_ok=True)


def frames(path):
    cap = cv2.VideoCapture(path)
    if not cap.isOpened():
        sys.exit(f"cannot open video: {path}")
    while True:
        ok, f = cap.read()
        if not ok:
            break
        yield cv2.cvtColor(f, cv2.COLOR_BGR2GRAY)
    cap.release()


def frame_to_text(mask, heat, pad_frac=0.08):
    # Crop both images tightly to the text and enlarge, so a small glyph (a lone
    # `I`, an accent, a short top line) is big and obvious instead of a few pixels
    # lost in a mostly-empty frame. The mask defines the box; the heatmap matches.
    ys, xs = np.where(mask > 127)
    if ys.size == 0:
        return mask, heat
    y0, y1, x0, x1 = int(ys.min()), int(ys.max()), int(xs.min()), int(xs.max())
    hh, ww = mask.shape
    pad = int(pad_frac * max(x1 - x0, y1 - y0)) + 8
    y0, y1 = max(0, y0 - pad), min(hh, y1 + pad + 1)
    x0, x1 = max(0, x0 - pad), min(ww, x1 + pad + 1)
    mask, heat = mask[y0:y1, x0:x1], heat[y0:y1, x0:x1]
    long_side = max(mask.shape[:2])
    if long_side < 1000:
        f = min(4.0, 1000.0 / long_side)
        size = (int(mask.shape[1] * f), int(mask.shape[0] * f))
        mask = cv2.resize(mask, size, interpolation=cv2.INTER_NEAREST)
        heat = cv2.resize(heat, size, interpolation=cv2.INTER_CUBIC)
    return mask, heat


# --- accumulate motion against the background (letters move, background drifts) ---
dis = cv2.DISOpticalFlow_create(cv2.DISOPTICAL_FLOW_PRESET_MEDIUM)
score = prev = prev_smooth = None
drift = np.zeros(2)
for gray in frames(VIDEO):
    if prev is not None:
        flow = dis.calc(prev, gray, None)
        bg = np.median(flow.reshape(-1, 2), axis=0)
        residual = flow - bg
        mag = float(np.hypot(*bg))
        ps = (residual @ (-bg / mag)) if mag > 0.15 else np.hypot(residual[..., 0], residual[..., 1])
        ps = np.clip(ps, 0, None).astype(np.float32)
        smooth = cv2.GaussianBlur(ps, (31, 31), 0)
        if prev_smooth is not None:
            (dx, dy), r = cv2.phaseCorrelate(prev_smooth, smooth)
            if r > 0.05 and np.hypot(dx, dy) < 30:
                drift += (dx, dy)
        prev_smooth = smooth
        h, w = ps.shape
        M = np.float32([[1, 0, -drift[0]], [0, 1, -drift[1]]])
        reg = cv2.warpAffine(ps, M, (w, h))
        score = reg if score is None else score + reg
    prev = gray
if score is None:
    sys.exit("fewer than 2 usable frames")

# --- build the two images: heatmap (raw score) + clean mask ---
score = np.clip(score, 0, None)
hi = np.percentile(score, 99.5)
norm = np.clip(score / hi * 255, 0, 255).astype(np.uint8) if hi > 0 else score.astype(np.uint8)
norm = cv2.GaussianBlur(norm, (5, 5), 0)

_, mask = cv2.threshold(norm, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
k = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (7, 7))
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, k)
mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, k)
h_img, w_img = mask.shape
n, lab, st, _ = cv2.connectedComponentsWithStats(mask)
for i in range(1, n):
    x, y, w, h, area = st[i]
    band = w >= 5 * h and h <= h_img // 18            # wide, short
    at_edge = x <= 2 or x + w >= w_img - 2            # drift bands hug an edge
    if area < mask.size // 20000 or (band and (at_edge or w >= w_img // 3)):
        mask[lab == i] = 0

# --- optional OCR hint on the full-frame mask (never authoritative) ---
try:
    import pytesseract
    exe = shutil.which("tesseract")
    if exe:
        pytesseract.pytesseract.tesseract_cmd = exe
        t = pytesseract.image_to_string(cv2.bitwise_not(mask), config="--psm 6").strip()
        print("OCR hint (unreliable):", " ".join(t.split()) if t else "(none)")
except Exception:
    pass

# --- crop both images tightly to the text (a lone I stays visible), then save ---
mask, norm = frame_to_text(mask, norm)
cv2.imwrite(os.path.join(OUT, "revealed_heatmap.png"), norm)
cv2.imwrite(os.path.join(OUT, "revealed.png"), mask)
print("done — wrote revealed.png and revealed_heatmap.png (the only two outputs)")