Security audit
Security checks across malware telemetry and agentic risk
The skill coherently analyzes Word, PDF, and PowerPoint documents, with expected but privacy-relevant image rendering and caption-helper processing.
Install only if you are comfortable with the agent reading the selected documents, rendering pages or embedded images to temporary local files, and passing those images to the referenced captioning helper. For sensitive business, legal, HR, or financial documents, confirm how the captioning helper processes data and clean temporary files after use.
else:
cmd += ["--prompt", "提取页面中所有文字和表格内容，保持原始结构，Markdown格式输出。"]
r = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
if r.returncode == 0:
desc = json.loads(r.stdout).get("description", "")
all_text.append(f"=== Page {i+1} ===\n{desc}")prompt = image_prompt or "提取页面中所有文字和表格内容，Markdown格式输出。"
cmd += ["--prompt", prompt]
r = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
if r.returncode == 0:
desc = json.loads(r.stdout).get("description", "")
all_text.append(f"=== Page {i+1} (image→caption) ===\n{desc}")cmd = ["python3", CAPTION, img_path, "--json"]
p = prompt or "提取幻灯片中所有文字、数值和表格内容，保持结构，Markdown格式输出。"
cmd += ["--prompt", p]
cr = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
if cr.returncode == 0:
desc = json.loads(cr.stdout).get("description", "")
s['text'] = desccmd = ["python3", CAPTION, tmp_path, "--json"]
if prompt:
cmd += ["--prompt", prompt]
r = subprocess.run(cmd, capture_output=True, text=True, timeout=60)
if r.returncode == 0:
desc = json.loads(r.stdout).get("description", "")
results.append({'image': media, 'caption': desc})60/60 vendors flagged this skill as clean.
No suspicious patterns detected.