qianwen-vision — Skillopedia

Qwen Vision (Image & Video Understanding) Analyze images and videos using Qwen VL and QVQ models. This skill is part of QianWen-AI/qianwen-ai . Skill directory Use this skill's internal files to execute and learn. Load reference files on demand when the default path fails or you need details. | Location | Purpose | |----------|---------| | | Image/video understanding, multi-image, thinking mode | | | Visual reasoning (QVQ, chain-of-thought, streaming) | | | OCR text extraction | | | Shared helpers (base64, upload, streaming) | | | Fallback: curl, code generation | | | Curl for base64, multi-i…

\n for candidate in candidates:\n m = re.search(pattern, candidate)\n if m:\n agent, skill = m.group(1) or \"default\", m.group(2)\n if len(agent) \u003c= 32 and len(skill) \u003c= 32:\n return json.dumps({\"channel\": \"qianwenai-skill\", \"tags\": {\"t1\": skill, \"t2\": agent}}, separators=(',', ':'))\n return None\n\n\n# ---------------------------------------------------------------------------\n# Section 6: HTTP infrastructure\n# ---------------------------------------------------------------------------\n\n_RETRYABLE_CODES = frozenset({429, 500, 502, 503, 504})\n\n\ndef http_request(\n method: str,\n url: str,\n api_key: str,\n payload: dict[str, Any] | None = None,\n *,\n extra_headers: dict[str, str] | None = None,\n timeout: int = 120,\n retries: int = 2,\n backoff: float = 1.5,\n) -> dict[str, Any]:\n \"\"\"Generic HTTP request with retry and exponential backoff.\n\n Handles JSON serialisation, provider header injection, and retryable\n HTTP codes.\n \"\"\"\n hdrs = get_provider().make_headers(api_key, payload)\n if extra_headers:\n hdrs.update(extra_headers)\n data = json.dumps(payload).encode(\"utf-8\") if payload else None\n last_err = \"\"\n for attempt in range(retries + 1):\n req = urllib.request.Request(url, data=data, headers=hdrs, method=method)\n try:\n with urllib.request.urlopen(req, timeout=timeout) as resp:\n return json.loads(resp.read().decode(\"utf-8\"))\n except urllib.error.HTTPError as exc:\n body = \"\"\n try:\n body = exc.read().decode(\"utf-8\", errors=\"replace\")\n except Exception:\n pass\n last_err = f\"HTTP {exc.code}: {body[:500]}\"\n if api_key and len(api_key) > 8 and api_key in body:\n last_err = last_err.replace(api_key, mask_key(api_key))\n if exc.code not in _RETRYABLE_CODES or attempt >= retries:\n raise RuntimeError(last_err) from exc\n time.sleep(backoff * (2 ** attempt))\n except urllib.error.URLError as exc:\n last_err = str(exc.reason)\n if attempt >= retries:\n raise RuntimeError(f\"Network error: {last_err}\") from exc\n time.sleep(backoff * (2 ** attempt))\n raise RuntimeError(last_err)\n\n\ndef http_post(\n url: str,\n api_key: str,\n payload: dict[str, Any],\n *,\n timeout: int = 120,\n retries: int = 2,\n backoff: float = 1.5,\n) -> dict[str, Any]:\n \"\"\"Convenience: non-streaming POST, returns parsed JSON.\"\"\"\n return http_request(\n \"POST\", url, api_key, payload,\n timeout=timeout, retries=retries, backoff=backoff,\n )\n\n\ndef _sse_error_message(\n chunk: Any,\n event_type: str,\n api_key: str,\n) -> Optional[str]:\n \"\"\"Return a safe error message for an SSE error payload, if present.\"\"\"\n error_payload = chunk.get(\"error\") if isinstance(chunk, dict) else None\n has_top_level_error = isinstance(chunk, dict) and (\n not chunk.get(\"choices\")\n and chunk.get(\"code\") is not None\n and chunk.get(\"message\") is not None\n )\n if event_type.lower() != \"error\" and error_payload is None and not has_top_level_error:\n return None\n\n payload = error_payload if error_payload is not None else chunk\n if isinstance(payload, dict):\n code = payload.get(\"code\") or payload.get(\"type\")\n message = payload.get(\"message\")\n if code and message:\n detail = f\"{code}: {message}\"\n elif code or message:\n detail = str(code or message)\n else:\n detail = json.dumps(payload, ensure_ascii=False)\n else:\n detail = str(payload)\n\n if api_key and len(api_key) > 8 and api_key in detail:\n detail = detail.replace(api_key, mask_key(api_key))\n return f\"SSE error: {sanitize_diagnostic(detail)[:500]}\"\n\n\ndef stream_sse(\n url: str,\n api_key: str,\n payload: dict[str, Any],\n *,\n timeout: int = 180,\n) -> Iterator[dict[str, Any]]:\n \"\"\"Streaming POST. Yields parsed SSE ``data:`` chunks.\n\n Automatically sets ``stream: true`` in the payload and sends the\n appropriate ``Accept`` header.\n \"\"\"\n payload[\"stream\"] = True\n hdrs = get_provider().make_headers(api_key, payload)\n hdrs[\"Accept\"] = \"text/event-stream\"\n data = json.dumps(payload).encode(\"utf-8\")\n req = urllib.request.Request(url, data=data, headers=hdrs, method=\"POST\")\n try:\n resp = urllib.request.urlopen(req, timeout=timeout)\n except urllib.error.HTTPError as exc:\n body = \"\"\n try:\n body = exc.read().decode(\"utf-8\", errors=\"replace\")\n except Exception:\n pass\n if api_key and len(api_key) > 8 and api_key in body:\n body = body.replace(api_key, mask_key(api_key))\n raise RuntimeError(f\"HTTP {exc.code}: {sanitize_diagnostic(body)[:500]}\") from None\n except urllib.error.URLError as exc:\n raise RuntimeError(f\"Network error: {sanitize_diagnostic(str(exc.reason))}\") from None\n\n buf = b\"\"\n event_type = \"\"\n try:\n while True:\n raw = resp.read(4096)\n if not raw:\n break\n buf += raw\n while b\"\\n\" in buf:\n line_bytes, buf = buf.split(b\"\\n\", 1)\n line = line_bytes.decode(\"utf-8\", errors=\"replace\").strip()\n if not line:\n event_type = \"\"\n continue\n if line.startswith(\"event:\"):\n event_type = line[6:].strip()\n continue\n if not line.startswith(\"data:\"):\n continue\n json_str = line[5:].strip()\n if json_str == \"[DONE]\":\n return\n try:\n chunk = json.loads(json_str)\n except json.JSONDecodeError:\n if event_type.lower() == \"error\":\n detail = json_str\n if api_key and len(api_key) > 8 and api_key in detail:\n detail = detail.replace(api_key, mask_key(api_key))\n detail = sanitize_diagnostic(detail)[:500]\n raise RuntimeError(f\"SSE error: {detail}\") from None\n pass\n else:\n error_message = _sse_error_message(chunk, event_type, api_key)\n if error_message:\n raise RuntimeError(error_message)\n yield chunk\n event_type = \"\"\n finally:\n resp.close()\n\n\n# ---------------------------------------------------------------------------\n# Section 7: File upload / download\n# ---------------------------------------------------------------------------\n\n_BASE64_FILE_LIMIT = 7 * 1024 * 1024 # 7 MB (base64 adds ~33%; API limit is 10 MB)\n\n\ndef upload_local_file(api_key: str, model: str, fp: Path) -> str:\n \"\"\"Upload a local file to the active provider's temp storage.\n\n Returns a provider-managed URL (e.g. ``oss://`` for DashScope).\n \"\"\"\n return get_provider().upload_file(api_key, model, fp)\n\n\ndef resolve_file(\n value: str,\n *,\n api_key: str | None = None,\n model: str | None = None,\n) -> str:\n \"\"\"Resolve a file reference for API consumption.\n\n URLs (``http``, ``https``, ``data``, and provider-managed schemes)\n pass through unchanged. Local files are handled based on context:\n\n - *api_key* + *model* provided: upload to temp storage.\n - Otherwise: convert to ``data:`` base64 URI (must be \u003c 7 MB).\n \"\"\"\n provider = get_provider()\n pass_through = (\"http://\", \"https://\", \"data:\") + provider.managed_url_schemes()\n if value.startswith(pass_through):\n return value\n if value.startswith(\"file://\"):\n parsed = urllib.parse.urlparse(value)\n if parsed.netloc and parsed.netloc.lower() != \"localhost\":\n raise ValueError(\n f\"Cannot resolve non-local file URI authority {parsed.netloc!r}; \"\n \"only an empty authority or 'localhost' is supported.\"\n )\n value = urllib.request.url2pathname(parsed.path)\n p = Path(value)\n if not (p.exists() and p.is_file()):\n return value\n\n file_size = p.stat().st_size\n\n if api_key and model:\n managed_url = provider.upload_file(api_key, model, p)\n tag = \"48 h TTL\" if managed_url.startswith(\"oss://\") else \"custom OSS\"\n print(sanitize_diagnostic(f\"Uploaded {p.name} -> {managed_url} ({tag})\"), file=sys.stderr)\n return managed_url\n\n if file_size > _BASE64_FILE_LIMIT:\n print(\n f\"Warning: {p.name} is {file_size / 1024 / 1024:.1f} MB -- \"\n \"base64 may exceed the 10 MB API limit. \"\n \"Use --upload-files to auto-upload, or provide an online URL.\",\n file=sys.stderr,\n )\n\n mime = mimetypes.guess_type(p.name)[0] or \"application/octet-stream\"\n b64 = base64.b64encode(p.read_bytes()).decode(\"ascii\")\n return f\"data:{mime};base64,{b64}\"\n\n\ndef download_file(url: str, dest: Path, *, timeout: int = 120) -> Path:\n \"\"\"Download a file from *url* to *dest*, creating parent dirs as needed.\"\"\"\n dest.parent.mkdir(parents=True, exist_ok=True)\n req = urllib.request.Request(url, headers={\"User-Agent\": \"qianwenai-skill\"})\n try:\n with urllib.request.urlopen(req, timeout=timeout) as resp:\n dest.write_bytes(resp.read())\n except urllib.error.URLError as exc:\n raise RuntimeError(f\"Download failed: {sanitize_diagnostic(str(exc))}\") from None\n return dest\n\n\n# ---------------------------------------------------------------------------\n# Section 8: Request / response utilities\n# ---------------------------------------------------------------------------\n\ndef load_request(args: Any) -> dict[str, Any]:\n \"\"\"Load request dict from ``--request`` or ``--file`` CLI argument.\"\"\"\n if getattr(args, \"request\", None):\n return json.loads(args.request)\n if getattr(args, \"file\", None):\n return json.loads(Path(args.file).read_text(encoding=\"utf-8\"))\n raise ValueError(\"Provide --request '{...}' or --file path/to/request.json\")\n\n\ndef save_result(result: dict[str, Any], output_path: str | Path) -> None:\n \"\"\"Write *result* as JSON to *output_path*, creating parent dirs.\"\"\"\n p = Path(output_path)\n p.parent.mkdir(parents=True, exist_ok=True)\n p.write_text(json.dumps(result, ensure_ascii=False, indent=2), encoding=\"utf-8\")\n\n\ndef extract_text(content: Any) -> str:\n \"\"\"Extract plain text from a message ``content`` field.\n\n Handles ``str``, ``list[{type, text}]``, and ``None``.\n \"\"\"\n if isinstance(content, str):\n return content\n if isinstance(content, list):\n for item in content:\n if isinstance(item, dict) and item.get(\"type\") == \"text\":\n text = item.get(\"text\")\n if isinstance(text, str):\n return text\n return json.dumps(content, ensure_ascii=False)\n if content is None:\n return \"\"\n return json.dumps(content, ensure_ascii=False)\n\n\ndef try_parse_json(text: str) -> Any | None:\n \"\"\"Attempt to parse *text* as JSON; return ``None`` on failure.\"\"\"\n text = text.strip()\n if not text:\n return None\n try:\n return json.loads(text)\n except json.JSONDecodeError:\n return None\n\n\n# ---------------------------------------------------------------------------\n# Section 9: Async task polling\n# ---------------------------------------------------------------------------\n\ndef poll_task(\n task_id: str,\n api_key: str,\n *,\n timeout_s: int = 600,\n interval: int = 10,\n verbose: bool = True,\n) -> dict[str, Any]:\n \"\"\"Poll an async task until it reaches a terminal state.\n\n Returns the final result dict. Raises ``TimeoutError`` if\n *timeout_s* is exceeded.\n \"\"\"\n provider = get_provider()\n url = provider.task_poll_url(task_id)\n terminals = provider.terminal_statuses()\n start = time.time()\n poll_count = 0\n while True:\n result = http_request(\"GET\", url, api_key)\n elapsed = int(time.time() - start)\n poll_count += 1\n status = provider.extract_task_status(result)\n if verbose:\n print(f\" [{elapsed}s] task={task_id} status={status}\", file=sys.stderr)\n if status in terminals:\n return result\n if time.time() - start > timeout_s:\n raise TimeoutError(\n f\"Task {task_id} timed out after {timeout_s}s \"\n f\"({poll_count} polls, last status: {status})\"\n )\n time.sleep(interval)\n\n\n# ---------------------------------------------------------------------------\n# Section 10: Update-check signal\n# ---------------------------------------------------------------------------\n\ntry:\n from gossamer import run as _run_update_signal_impl\nexcept ImportError:\n _run_update_signal_impl = None # type: ignore[assignment]\n\n\ndef run_update_signal(caller: str | Path | None = None) -> None:\n \"\"\"Emit update-check signals to stderr (non-blocking, failure-safe).\n\n Parameters\n ----------\n caller : str or Path, optional\n ``__file__`` of the calling script.\n \"\"\"\n if _run_update_signal_impl:\n try:\n _run_update_signal_impl(caller=caller or __file__)\n except Exception:\n pass\n","content_type":"text/x-python; charset=utf-8","language":"python","size":43758,"content_sha256":"9343069363f8e31b7178d2462c79e2e7d56cb7a30665544d9072b1d39589d8b0"},{"filename":"scripts/reason.py","content":"#!/usr/bin/env python3\n\"\"\"Visual reasoning with QVQ and thinking-enabled VL models.\n\nChain-of-thought visual reasoning: the model thinks step-by-step before answering.\nBest for math problems in images, complex chart analysis, video understanding,\nand multi-step visual logic. Always uses streaming (required for QVQ models).\n\nUsage:\n python scripts/reason.py --request '{\"prompt\":\"Solve this problem\",\"image\":\"math.png\"}'\n python scripts/reason.py --request '{\"prompt\":\"Analyze trends\",\"image\":\"chart.jpg\",\"model\":\"qwen3-vl-plus\"}'\n python scripts/reason.py --request '{\"prompt\":\"What happens?\",\"video\":\"clip.mp4\",\"fps\":2}'\n python scripts/reason.py --request '{\"prompt\":\"Describe\",\"video_frames\":[\"f1.jpg\",\"f2.jpg\"],\"fps\":2}'\n python scripts/reason.py --file request.json --output result.json\n\"\"\"\nfrom __future__ import annotations\n\nimport argparse\nimport json\nimport sys\nfrom pathlib import Path\nfrom typing import Any\n\nsys.path.insert(0, str(Path(__file__).resolve().parent))\n\nfrom vision_lib import ( # noqa: E402\n chat_url,\n get_default_model,\n load_request,\n prompt_update_check_install,\n require_api_key,\n save_result,\n stream_sse,\n build_content,\n validate_token_plan_model,\n)\n\nDEFAULT_MAX_TOKENS = 8192\n\ndef reason(req: dict[str, Any], api_key: str, *, upload_files: bool = False) -> dict[str, Any]:\n prompt = req.get(\"prompt\")\n if not prompt:\n raise ValueError(\"prompt is required\")\n\n model = req[\"model\"] if \"model\" in req else get_default_model(\"reason\")\n upload_key = api_key if upload_files else None\n upload_model = model if upload_files else None\n\n content = build_content(req, upload_key=upload_key, upload_model=upload_model)\n content.append({\"type\": \"text\", \"text\": prompt})\n\n payload: dict[str, Any] = {\n \"model\": model,\n \"messages\": [{\"role\": \"user\", \"content\": content}],\n \"max_tokens\": req.get(\"max_tokens\", DEFAULT_MAX_TOKENS),\n \"stream\": True,\n \"stream_options\": {\"include_usage\": True},\n }\n\n enable_thinking = req.get(\"enable_thinking\")\n if enable_thinking is not None:\n payload[\"enable_thinking\"] = bool(enable_thinking)\n thinking_budget = req.get(\"thinking_budget\")\n if thinking_budget is not None:\n payload[\"thinking_budget\"] = int(thinking_budget)\n\n reasoning_content = \"\"\n answer_content = \"\"\n usage: dict[str, Any] = {}\n model_name = model\n is_answering = False\n\n for chunk in stream_sse(\n chat_url(), api_key, payload,\n timeout=int(req.get(\"timeout_s\", 300)),\n ):\n if chunk.get(\"usage\"):\n usage = chunk[\"usage\"]\n if chunk.get(\"model\"):\n model_name = chunk[\"model\"]\n\n choices = chunk.get(\"choices\") or []\n if not choices:\n continue\n delta = choices[0].get(\"delta\") or {}\n\n rc = delta.get(\"reasoning_content\") or \"\"\n if rc:\n reasoning_content += rc\n if not is_answering:\n print(rc, end=\"\", flush=True, file=sys.stderr)\n\n ct = delta.get(\"content\") or \"\"\n if ct:\n if not is_answering:\n is_answering = True\n if reasoning_content:\n print(\"\\n\", file=sys.stderr)\n answer_content += ct\n print(ct, end=\"\", flush=True, file=sys.stderr)\n\n if reasoning_content or answer_content:\n print(\"\", file=sys.stderr)\n\n result: dict[str, Any] = {\n \"text\": answer_content,\n \"model\": model_name,\n \"usage\": usage,\n }\n if reasoning_content:\n result[\"reasoning\"] = reasoning_content\n return result\n\ndef main() -> None:\n prompt_update_check_install()\n help_default_model = (\n get_default_model(\"reason\")\n if any(arg in (\"-h\", \"--help\") for arg in sys.argv[1:])\n else \"loaded from CDN/fallback configuration\"\n )\n parser = argparse.ArgumentParser(\n description=\"Visual reasoning with QVQ/thinking VL models (always streaming)\",\n formatter_class=argparse.RawDescriptionHelpFormatter,\n epilog=\"\"\"\\\nrequest JSON fields (--request / --file):\n prompt (required) Reasoning question or instruction\n image Single image URL or local path\n video Video URL or local path\n video_frames Array of frame image URLs/paths (alternative to video)\n fps Frame sampling rate for video (default: auto)\n model Model ID (default: {help_default_model})\n enable_thinking true/false — override thinking mode (QVQ always thinks)\n thinking_budget Max thinking tokens\n max_tokens Max output tokens (default: 8192)\n timeout_s Request timeout in seconds (default: 300)\n\ninput types (provide exactly one):\n image Single image reasoning\n video Video reasoning (URL or local file)\n video_frames Video from extracted frames (array of image paths)\n\nmodels:\n qvq-max Visual reasoning specialist (always-on thinking)\n qwen3-vl-plus General vision with optional thinking mode\n qwen3.5-plus Unified multimodal with thinking on by default\n\noutput:\n Result JSON contains: text (answer), reasoning (chain-of-thought), usage\n\nenvironment variables:\n DASHSCOPE_API_KEY (required) API key — also loaded from .env\n QIANWEN_API_KEY (alternative) Alias for DASHSCOPE_API_KEY\n QWEN_REGION cn-beijing (default)\n\nexamples:\n # Solve a math problem from image\n python scripts/reason.py --request '{\"prompt\":\"Solve this\",\"image\":\"math.png\"}'\n\n # Chart analysis with thinking\n python scripts/reason.py --request '{\"prompt\":\"Analyze trends\",\"image\":\"chart.jpg\",\n \"model\":\"qwen3-vl-plus\",\"enable_thinking\":true}' --output result.json\n\n # Video reasoning\n python scripts/reason.py --request '{\"prompt\":\"What happens and why?\",\n \"video\":\"clip.mp4\",\"fps\":2}' --print-response\n\"\"\".replace(\"{help_default_model}\", help_default_model),\n )\n parser.add_argument(\"--request\", help=\"Inline JSON: must contain 'prompt' + image/video input\")\n parser.add_argument(\"--file\", help=\"Path to JSON request file\")\n parser.add_argument(\"--model\", type=str, default=None,\n help=f\"Model ID (default: {help_default_model})\")\n parser.add_argument(\"--upload-files\", action=\"store_true\",\n help=\"Upload local files to temp storage (oss://) instead of base64\")\n parser.add_argument(\"--output\", default=\"\", help=\"Save result JSON to this path\")\n parser.add_argument(\"--print-response\", action=\"store_true\", help=\"Print result JSON to stdout\")\n args = parser.parse_args()\n\n api_key = require_api_key()\n\n try:\n req = load_request(args)\n except ValueError as e:\n print(f\"Error: {e}\", file=sys.stderr)\n sys.exit(1)\n\n if args.model:\n req[\"model\"] = args.model\n elif \"model\" not in req or not req.get(\"model\"):\n req[\"model\"] = get_default_model(\"reason\")\n\n validate_token_plan_model(api_key, req[\"model\"])\n\n try:\n result = reason(req, api_key, upload_files=args.upload_files)\n except (ValueError, RuntimeError) as e:\n print(f\"Error: {e}\", file=sys.stderr)\n sys.exit(1)\n\n if args.output:\n save_result(result, args.output)\n if args.print_response:\n print(json.dumps(result, ensure_ascii=False, indent=2))\n\nif __name__ == \"__main__\":\n main()\n","content_type":"text/x-python; charset=utf-8","language":"python","size":7376,"content_sha256":"96294b0e608bed0c56f8c5a13ae003bba12ee2e57904a7d9fddfa6b451386f84"},{"filename":"scripts/vision_lib.py","content":"\"\"\"Shared infrastructure for qianwen-vision scripts.\n\nThin wrapper around qianwen_lib that re-exports common symbols and adds\nvision-specific helpers (streaming with reasoning content, base64 encoding for\nsmaller files, shared multimodal content builder). Stdlib only -- no pip install\nrequired.\n\"\"\"\nfrom __future__ import annotations\n\nimport sys\n\nif sys.version_info \u003c (3, 9):\n print(f\"Error: Python 3.9+ required (found {sys.version}). \"\n \"Install: https://www.python.org/downloads/\", file=sys.stderr)\n sys.exit(1)\n\nfrom pathlib import Path\nfrom typing import Any\n\nsys.path.insert(0, str(Path(__file__).resolve().parent))\n\n# Re-export shared infrastructure so existing scripts keep working with\n# ``from vision_lib import ...``.\nfrom qianwen_lib import ( # noqa: E402,F401\n chat_url,\n extract_text,\n has_oss_url,\n http_post,\n load_cdn_model_config,\n load_request,\n native_base_url,\n require_api_key as _require_api_key_base,\n resolve_file,\n run_update_signal,\n save_result,\n stream_sse,\n try_parse_json,\n upload_local_file,\n validate_token_plan_model,\n)\n\n\ndef get_default_model(task: str) -> str:\n \"\"\"Return the current default model for a vision task from the CDN config.\"\"\"\n config = load_cdn_model_config(\n \"qianwen-vision-config.json\",\n required_keys=(\"default_models\",),\n )\n defaults = config[\"default_models\"]\n if not isinstance(defaults, dict):\n raise RuntimeError(\"Invalid vision model configuration: default_models must be an object\")\n model = defaults.get(task)\n if not isinstance(model, str) or not model:\n raise RuntimeError(f\"Invalid vision model configuration: missing default for {task}\")\n return model\n\n\n# ---------------------------------------------------------------------------\n# Vision-specific credential wrapper\n# ---------------------------------------------------------------------------\n\ndef require_api_key() -> str:\n \"\"\"Load API key with Vision domain tagging.\"\"\"\n return _require_api_key_base(script_file=__file__, domain=\"Vision\")\n\n# ---------------------------------------------------------------------------\n# Update-check signal (convenience)\n# ---------------------------------------------------------------------------\n\ndef prompt_update_check_install() -> None:\n \"\"\"Emit structured signals about update-check skill status to stderr.\"\"\"\n run_update_signal(caller=__file__)\n\n# ---------------------------------------------------------------------------\n# Shared multimodal content builder\n# ---------------------------------------------------------------------------\n\ndef build_content(\n req: dict[str, Any],\n detail: str = \"auto\",\n *,\n upload_key: str | None = None,\n upload_model: str | None = None,\n) -> list[dict[str, Any]]:\n \"\"\"Build multimodal content parts from image/video/video_frames input.\n\n Shared by analyze.py and reason.py. The ``detail`` parameter is only\n applied to image_url items and is ignored for video content.\n\n Supports:\n - ``video_frames``: list of frame image paths/URLs → video content part\n - ``video``: single video URL/path → video_url content part\n - ``images``: list of image paths/URLs → multiple image_url parts\n - ``image``: single image path/URL → one image_url part\n \"\"\"\n parts: list[dict[str, Any]] = []\n\n if req.get(\"video_frames\"):\n frames = [\n resolve_file(str(frame), api_key=upload_key, model=upload_model)\n for frame in req[\"video_frames\"]\n ]\n video_obj: dict[str, Any] = {\"type\": \"video\", \"video\": frames}\n if req.get(\"fps\"):\n video_obj[\"fps\"] = float(req[\"fps\"])\n parts.append(video_obj)\n elif req.get(\"video\"):\n video_url = resolve_file(str(req[\"video\"]), api_key=upload_key, model=upload_model)\n video_obj = {\n \"type\": \"video_url\",\n \"video_url\": {\"url\": video_url},\n }\n if req.get(\"fps\"):\n video_obj[\"fps\"] = float(req[\"fps\"])\n parts.append(video_obj)\n elif req.get(\"images\"):\n for img in req[\"images\"]:\n resolved_url = resolve_file(str(img), api_key=upload_key, model=upload_model)\n parts.append({\"type\": \"image_url\", \"image_url\": {\"url\": resolved_url, \"detail\": detail}})\n elif req.get(\"image\"):\n resolved_url = resolve_file(str(req[\"image\"]), api_key=upload_key, model=upload_model)\n parts.append({\"type\": \"image_url\", \"image_url\": {\"url\": resolved_url, \"detail\": detail}})\n else:\n raise ValueError(\"Provide 'image', 'images', 'video', or 'video_frames'\")\n\n return parts\n","content_type":"text/x-python; charset=utf-8","language":"python","size":4649,"content_sha256":"e75ea5b4b96dd180cc9d2c652f7282fa5097b8b871df771953383391f2b6baec"}],"content_json":{"type":"doc","content":[{"type":"heading","attrs":{"level":1},"content":[{"text":"Qwen Vision (Image & Video Understanding)","type":"text"}]},{"type":"paragraph","content":[{"text":"Analyze images and videos using Qwen VL and QVQ models. This skill is part of ","type":"text"},{"text":"QianWen-AI/qianwen-ai","type":"text","marks":[{"type":"strong"}]},{"text":".","type":"text"}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Skill directory","type":"text"}]},{"type":"paragraph","content":[{"text":"Use this skill's internal files to execute and learn. Load reference files on demand when the default path fails or you need details.","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Location","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Purpose","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/analyze.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Image/video understanding, multi-image, thinking mode","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/reason.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Visual reasoning (QVQ, chain-of-thought, streaming)","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/ocr.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"OCR text extraction","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/vision_lib.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Shared helpers (base64, upload, streaming)","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"references/execution-guide.md","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Fallback: curl, code generation","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"references/curl-examples.md","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Curl for base64, multi-image, video, OCR","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"references/visual-reasoning.md","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"QVQ and thinking mode details","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"references/prompt-guide.md","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Query prompt templates by task, thinking mode decision","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"references/ocr.md","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"OCR parameters and examples","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"references/sources.md","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Official documentation URLs","type":"text"}]}]}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Security","type":"text"}]},{"type":"paragraph","content":[{"text":"NEVER output any API key or credential in plaintext.","type":"text","marks":[{"type":"strong"}]},{"text":" Always use variable references (","type":"text"},{"text":"$DASHSCOPE_API_KEY","type":"text","marks":[{"type":"code_inline"}]},{"text":" in shell, ","type":"text"},{"text":"os.environ[\"DASHSCOPE_API_KEY\"]","type":"text","marks":[{"type":"code_inline"}]},{"text":" in Python). Any check or detection of credentials must be ","type":"text"},{"text":"non-plaintext","type":"text","marks":[{"type":"strong"}]},{"text":": report only status (e.g. \"set\" / \"not set\", \"valid\" / \"invalid\"), never the value. Never display contents of ","type":"text"},{"text":".env","type":"text","marks":[{"type":"code_inline"}]},{"text":" or config files that may contain secrets.","type":"text"}]},{"type":"paragraph","content":[{"text":"When the API key is not configured, NEVER ask the user to provide it directly.","type":"text","marks":[{"type":"strong"}]},{"text":" Instead, help create a ","type":"text"},{"text":".env","type":"text","marks":[{"type":"code_inline"}]},{"text":" file with a placeholder (","type":"text"},{"text":"DASHSCOPE_API_KEY=sk-your-key-here","type":"text","marks":[{"type":"code_inline"}]},{"text":") and instruct the user to replace it with their actual key from the ","type":"text"},{"text":"QianWen Console","type":"text","marks":[{"type":"link","attrs":{"href":"https://platform.qianwenai.com/home/api-keys","title":null}}]},{"text":". Only write the actual key value if the user explicitly requests it.","type":"text"}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Key Compatibility","type":"text"}]},{"type":"paragraph","content":[{"text":"Both PAYG (","type":"text"},{"text":"sk-ws-...","type":"text","marks":[{"type":"code_inline"}]},{"text":"; legacy ","type":"text"},{"text":"sk-...","type":"text","marks":[{"type":"code_inline"}]},{"text":") and Token Plan (","type":"text"},{"text":"sk-sp-...","type":"text","marks":[{"type":"code_inline"}]},{"text":") keys are supported. Detect the API key type without exposing the Key:","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":"bash"},"content":[{"text":"python3 -c \"\nimport sys; sys.path.insert(0, 'scripts')\nfrom qianwen_lib import detect_api_key_type\nprint(detect_api_key_type('scripts/qianwen_lib.py'))\n\"","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Output","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Meaning","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"token-plan","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Token Plan key — use only models from the Token Plan catalog below.","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"payg","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Pay-as-you-go key — full model catalog available.","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"not-set","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"No key configured.","type":"text"}]}]}]}]},{"type":"paragraph","content":[{"text":"For Token Plan, fetch and read the current ","type":"text"},{"text":"Token Plan model catalog","type":"text","marks":[{"type":"link","attrs":{"href":"https://alioth.alicdn.com/skills-info/models/references/qianwen-token-plan-models.md","title":null}}]},{"text":", then use an exact listed vision-capable model. If CDN access fails, use the ","type":"text"},{"text":"local fallback","type":"text","marks":[{"type":"link","attrs":{"href":"cdn/references/qianwen-token-plan-models.md","title":null}}]},{"text":".","type":"text"}]},{"type":"paragraph","content":[{"text":"Token Plan supports only specific models — use exactly a model from the references above; do not guess or probe model availability. For PAYG, continue below.","type":"text"}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Token Plan users","type":"text","marks":[{"type":"strong"}]},{"text":": Use the Qwen vision model catalog below for current compatible substitutes when a task default is not on Token Plan.","type":"text"}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Model Selection","type":"text"}]},{"type":"paragraph","content":[{"text":"Before selecting, recommending, or defaulting a model, fetch and read the current ","type":"text"},{"text":"Qwen vision model catalog","type":"text","marks":[{"type":"link","attrs":{"href":"https://alioth.alicdn.com/skills-info/models/references/qianwen-vision-models.md","title":null}}]},{"text":". It contains the model list, basic model information, task recommendations, compatibility notes, and defaults. If CDN access fails, use the ","type":"text"},{"text":"local fallback","type":"text","marks":[{"type":"link","attrs":{"href":"cdn/references/qianwen-vision-models.md","title":null}}]},{"text":".","type":"text"}]},{"type":"ordered_list","attrs":{"order":1,"listStyle":"number"},"content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"User specified a model","type":"text","marks":[{"type":"strong"}]},{"text":" → use directly.","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Consult the qianwen-model-selector skill","type":"text","marks":[{"type":"strong"}]},{"text":" when model choice depends on requirement, scenario, or pricing.","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"No signal, clear task","type":"text","marks":[{"type":"strong"}]},{"text":" → use the default and task-specific alternatives from the model catalog.","type":"text"}]}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"⚠️ Important","type":"text","marks":[{"type":"strong"}]},{"text":": The model catalog is a ","type":"text"},{"text":"point-in-time snapshot","type":"text","marks":[{"type":"strong"}]},{"text":" and may be outdated. Model availability changes frequently. ","type":"text"},{"text":"Always check the ","type":"text","marks":[{"type":"strong"}]},{"text":"official model list","type":"text","marks":[{"type":"link","attrs":{"href":"https://www.qianwenai.com/models","title":null}},{"type":"strong"}]},{"text":" for the authoritative, up-to-date catalog before making model decisions.","type":"text","marks":[{"type":"strong"}]}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Model details","type":"text","marks":[{"type":"strong"}]},{"text":": For more information about a specific model, direct the user to ","type":"text"},{"text":"https://www.qianwenai.com/models/\u003cmodel-name>","type":"text","marks":[{"type":"code_inline"}]},{"text":". Replace ","type":"text"},{"text":"\u003cmodel-name>","type":"text","marks":[{"type":"code_inline"}]},{"text":" with the exact model ID; never modify or guess it.","type":"text"}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Dynamic model queries","type":"text","marks":[{"type":"strong"}]},{"text":": If the ","type":"text"},{"text":"qianwen-model-selector","type":"text","marks":[{"type":"strong"}]},{"text":" skill or ","type":"text"},{"text":"QianWen CLI","type":"text","marks":[{"type":"strong"}]},{"text":" (","type":"text"},{"text":"qianwen models info \u003cmodel>","type":"text","marks":[{"type":"code_inline"}]},{"text":") is available, use it for real-time model data. CLI requires authentication — see the ","type":"text"},{"text":"qianwen-usage","type":"text","marks":[{"type":"strong"}]},{"text":" skill for login flow.","type":"text"}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Execution","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Prerequisites","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"API Key","type":"text","marks":[{"type":"strong"}]},{"text":": Use the non-plaintext detector in ","type":"text"},{"text":"Key Compatibility","type":"text","marks":[{"type":"strong"}]},{"text":"; do not replace it with a variable-presence check. If no Key is found, use qianwen-ops-auth when available or guide the user to configure ","type":"text"},{"text":"DASHSCOPE_API_KEY","type":"text","marks":[{"type":"code_inline"}]},{"text":"/","type":"text"},{"text":"QIANWEN_API_KEY","type":"text","marks":[{"type":"code_inline"}]},{"text":" in ","type":"text"},{"text":".env","type":"text","marks":[{"type":"code_inline"}]},{"text":". Skills may be installed independently.","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Python 3.9+ (stdlib only, ","type":"text"},{"text":"no pip install needed","type":"text","marks":[{"type":"strong"}]},{"text":")","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Environment Check","type":"text"}]},{"type":"paragraph","content":[{"text":"Before first execution, verify Python is available:","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":"bash"},"content":[{"text":"python3 --version # must be 3.9+","type":"text"}]},{"type":"paragraph","content":[{"text":"If ","type":"text"},{"text":"python3","type":"text","marks":[{"type":"code_inline"}]},{"text":" is not found, try ","type":"text"},{"text":"python --version","type":"text","marks":[{"type":"code_inline"}]},{"text":" or ","type":"text"},{"text":"py -3 --version","type":"text","marks":[{"type":"code_inline"}]},{"text":". If Python is unavailable or below 3.9, skip to ","type":"text"},{"text":"Path 2 (curl)","type":"text","marks":[{"type":"strong"}]},{"text":" in ","type":"text"},{"text":"execution-guide.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/execution-guide.md","title":null}}]},{"text":".","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Default: Run Script","type":"text"}]},{"type":"paragraph","content":[{"text":"Script path","type":"text","marks":[{"type":"strong"}]},{"text":": Scripts are in the ","type":"text"},{"text":"scripts/","type":"text","marks":[{"type":"code_inline"}]},{"text":" subdirectory ","type":"text"},{"text":"of this skill's directory","type":"text","marks":[{"type":"strong"}]},{"text":" (the directory containing this SKILL.md). ","type":"text"},{"text":"You MUST first locate this skill's installation directory, then ALWAYS use the full absolute path to execute scripts.","type":"text","marks":[{"type":"strong"}]},{"text":" Do NOT assume scripts are in the current working directory. Do NOT use ","type":"text"},{"text":"cd","type":"text","marks":[{"type":"code_inline"}]},{"text":" to switch directories before execution. Shared infrastructure lives in ","type":"text"},{"text":"scripts/vision_lib.py","type":"text","marks":[{"type":"code_inline"}]},{"text":".","type":"text"}]},{"type":"paragraph","content":[{"text":"Execution note:","type":"text","marks":[{"type":"strong"}]},{"text":" Run all scripts in the ","type":"text"},{"text":"foreground","type":"text","marks":[{"type":"strong"}]},{"text":" — wait for stdout; do not background.","type":"text"}]},{"type":"paragraph","content":[{"text":"Discovery:","type":"text","marks":[{"type":"strong"}]},{"text":" Run ","type":"text"},{"text":"python3 \u003cthis-skill-dir>/scripts/analyze.py --help","type":"text","marks":[{"type":"code_inline"}]},{"text":" (or ","type":"text"},{"text":"reason.py","type":"text","marks":[{"type":"code_inline"}]},{"text":", ","type":"text"},{"text":"ocr.py","type":"text","marks":[{"type":"code_inline"}]},{"text":") first to see all available arguments.","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Script","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Purpose","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Default Model","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/analyze.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Image understanding, multi-image, video, thinking mode, high-res","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Read the current default from the model catalog above","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/reason.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Visual reasoning with chain-of-thought, video reasoning (always streaming)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Read the current default from the model catalog above","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"scripts/ocr.py","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"OCR text extraction from documents, receipts, tables","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Read the current default from the model catalog above","type":"text"}]}]}]}]},{"type":"paragraph","content":[{"text":"Input type fields","type":"text","marks":[{"type":"strong"}]},{"text":" (use exactly one in ","type":"text"},{"text":"--request","type":"text","marks":[{"type":"code_inline"}]},{"text":" JSON):","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Field","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Use for","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Example","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"image\"","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Single image (URL or local path)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"image\": \"photo.jpg\"","type":"text","marks":[{"type":"code_inline"}]}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"images\"","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Multi-image comparison (array)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"images\": [\"a.jpg\", \"b.jpg\"]","type":"text","marks":[{"type":"code_inline"}]}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"video\"","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Video file (URL or local path)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"video\": \"clip.mp4\"","type":"text","marks":[{"type":"code_inline"}]}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"video_frames\"","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Video as frame array","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"\"video_frames\": [\"f1.jpg\", \"f2.jpg\"]","type":"text","marks":[{"type":"code_inline"}]}]}]}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"⚠️ Common mistake","type":"text","marks":[{"type":"strong"}]},{"text":": Do NOT use ","type":"text"},{"text":"\"image\"","type":"text","marks":[{"type":"code_inline"}]},{"text":" for video files — use ","type":"text"},{"text":"\"video\"","type":"text","marks":[{"type":"code_inline"}]},{"text":" instead.","type":"text"}]}]},{"type":"code_block","attrs":{"wrap":false,"language":"bash"},"content":[{"text":"# Image analysis\npython3 \u003cthis-skill-dir>/scripts/analyze.py \\\n --request '{\"prompt\":\"What is in this image?\",\"image\":\"https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20241022/emyrja/dog_and_girl.jpeg\"}' \\\n --output output/qianwen-vision/result.json --print-response\n\n# Video analysis (local file — add --upload-files for files >= 7 MB)\npython3 \u003cthis-skill-dir>/scripts/analyze.py \\\n --request '{\"prompt\":\"Describe what happens in this video\",\"video\":\"clip.mp4\"}' \\\n --upload-files --print-response\n\npython3 \u003cthis-skill-dir>/scripts/reason.py \\\n --request '{\"prompt\":\"Solve this math problem step by step\",\"image\":\"problem.png\"}' \\\n --print-response\n\npython3 \u003cthis-skill-dir>/scripts/ocr.py \\\n --request '{\"image\":\"invoice.jpg\"}' \\\n --print-response","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Argument","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Description","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--request '{...}'","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"JSON request body","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--file path.json","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Load request from file","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--output path","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Save response JSON to path","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--print-response","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Print response to stdout","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--stream","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Enable streaming (auto for thinking/QVQ)","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--upload-files","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Upload local files to temp storage (for files > 7 MB)","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--model ID","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Override model (default per script: see table above)","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"--schema path.json","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"JSON Schema for structured extraction","type":"text"}]}]}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Model priority","type":"text","marks":[{"type":"strong"}]},{"text":": ","type":"text"},{"text":"--model","type":"text","marks":[{"type":"code_inline"}]},{"text":" CLI flag > ","type":"text"},{"text":"\"model\"","type":"text","marks":[{"type":"code_inline"}]},{"text":" field in ","type":"text"},{"text":"--request","type":"text","marks":[{"type":"code_inline"}]},{"text":" JSON > built-in default.","type":"text"}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Verify Result","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Exit code ","type":"text"},{"text":"0","type":"text","marks":[{"type":"code_inline"}]},{"text":" + output contains valid JSON with ","type":"text"},{"text":"choices","type":"text","marks":[{"type":"code_inline"}]},{"text":" field → ","type":"text"},{"text":"success","type":"text","marks":[{"type":"strong"}]}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Non-zero exit, HTTP error, empty response, or JSON with ","type":"text"},{"text":"\"code\"","type":"text","marks":[{"type":"code_inline"}]},{"text":"/","type":"text"},{"text":"\"message\"","type":"text","marks":[{"type":"code_inline"}]},{"text":" error → ","type":"text"},{"text":"fail","type":"text","marks":[{"type":"strong"}]}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Post-execution check","type":"text","marks":[{"type":"strong"}]},{"text":": When ","type":"text"},{"text":"--output","type":"text","marks":[{"type":"code_inline"}]},{"text":" is used, verify the response JSON file exists and contains expected content","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"MANDATORY — stderr signal check","type":"text","marks":[{"type":"strong"}]},{"text":": After confirming the result, scan the command's stderr output for ","type":"text"},{"text":"[ACTION_REQUIRED]","type":"text","marks":[{"type":"code_inline"}]},{"text":" or ","type":"text"},{"text":"[UPDATE_AVAILABLE]","type":"text","marks":[{"type":"code_inline"}]},{"text":". If either signal is present, you ","type":"text"},{"text":"MUST","type":"text","marks":[{"type":"strong"}]},{"text":" follow the instructions in ","type":"text"},{"text":"Update Check","type":"text","marks":[{"type":"link","attrs":{"href":"#update-check-mandatory-post-execution","title":null}}]},{"text":" below before responding to the user.","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"On Failure","type":"text"}]},{"type":"paragraph","content":[{"text":"If scripts fail, match the error output against the diagnostic table below to determine the resolution. If no match, read ","type":"text"},{"text":"execution-guide.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/execution-guide.md","title":null}}]},{"text":" for alternative paths: curl commands (Path 2), code generation (Path 3), and autonomous resolution (Path 5).","type":"text"}]},{"type":"paragraph","content":[{"text":"If Python is not available at all","type":"text","marks":[{"type":"strong"}]},{"text":" → skip directly to Path 2 (curl) in ","type":"text"},{"text":"execution-guide.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/execution-guide.md","title":null}}]},{"text":".","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Error Pattern","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Diagnosis","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Resolution","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"command not found: python3","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Python not on PATH","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Try ","type":"text"},{"text":"python","type":"text","marks":[{"type":"code_inline"}]},{"text":" or ","type":"text"},{"text":"py -3","type":"text","marks":[{"type":"code_inline"}]},{"text":"; install Python 3.9+ if missing","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Python 3.9+ required","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Script version check failed","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Upgrade Python to 3.9+","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"SyntaxError","type":"text","marks":[{"type":"code_inline"}]},{"text":" near type hints","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Python \u003c 3.9","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Upgrade Python to 3.9+","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"QIANWEN_API_KEY/DASHSCOPE_API_KEY not found","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Missing API key","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Obtain key from ","type":"text"},{"text":"QianWen Console","type":"text","marks":[{"type":"link","attrs":{"href":"https://platform.qianwenai.com/home/api-keys","title":null}}]},{"text":"; add to ","type":"text"},{"text":".env","type":"text","marks":[{"type":"code_inline"}]},{"text":": ","type":"text"},{"text":"echo 'DASHSCOPE_API_KEY=sk-...' >> .env","type":"text","marks":[{"type":"code_inline"}]},{"text":"; or run ","type":"text"},{"text":"qianwen-ops-auth","type":"text","marks":[{"type":"strong"}]},{"text":" if available","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"HTTP 401","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Invalid or mismatched key","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Run ","type":"text"},{"text":"qianwen-ops-auth","type":"text","marks":[{"type":"strong"}]},{"text":" (non-plaintext check only); verify key is valid","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"SSL: CERTIFICATE_VERIFY_FAILED","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"SSL cert issue (proxy/corporate)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"macOS: run ","type":"text"},{"text":"Install Certificates.command","type":"text","marks":[{"type":"code_inline"}]},{"text":"; else set ","type":"text"},{"text":"SSL_CERT_FILE","type":"text","marks":[{"type":"code_inline"}]},{"text":" env var","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"URLError","type":"text","marks":[{"type":"code_inline"}]},{"text":" / ","type":"text"},{"text":"ConnectionError","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Network unreachable","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Check internet; set ","type":"text"},{"text":"HTTPS_PROXY","type":"text","marks":[{"type":"code_inline"}]},{"text":" if behind proxy","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"HTTP 429","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Rate limited","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Wait and retry with backoff","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"HTTP 5xx","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Server error","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Retry with backoff","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"PermissionError","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Can't write output","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Use ","type":"text"},{"text":"--output","type":"text","marks":[{"type":"code_inline"}]},{"text":" to specify writable directory","type":"text"}]}]}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"File Input","type":"text"}]},{"type":"paragraph","content":[{"text":"The API accepts: ","type":"text"},{"text":"HTTP/HTTPS URL","type":"text","marks":[{"type":"strong"}]},{"text":", ","type":"text"},{"text":"Base64 data URI","type":"text","marks":[{"type":"strong"}]},{"text":", and ","type":"text"},{"text":"oss://","type":"text","marks":[{"type":"code_inline"},{"type":"strong"}]},{"text":" URL","type":"text","marks":[{"type":"strong"}]},{"text":". Local file paths are NOT directly supported — scripts handle conversion automatically. ","type":"text"},{"text":"Pass local paths directly; no manual upload step needed.","type":"text","marks":[{"type":"strong"}]}]},{"type":"paragraph","content":[{"text":"Large file rule: If the local file is >= 7 MB, always add ","type":"text","marks":[{"type":"strong"}]},{"text":"--upload-files","type":"text","marks":[{"type":"code_inline"},{"type":"strong"}]},{"text":".","type":"text","marks":[{"type":"strong"}]},{"text":" Base64 encoding inflates size by ~33% and will exceed the 10 MB API limit. Small files (including short video clips \u003c 7 MB) can use the default base64 path.","type":"text"}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Token Plan (","type":"text","marks":[{"type":"strong"}]},{"text":"sk-sp-...","type":"text","marks":[{"type":"code_inline"},{"type":"strong"}]},{"text":") exception","type":"text","marks":[{"type":"strong"}]},{"text":": The ","type":"text"},{"text":"--upload-files","type":"text","marks":[{"type":"code_inline"}]},{"text":" temp-upload endpoint is unavailable for Token Plan keys and will be blocked with a friendly error. Token Plan users should use the default base64 path for local files; for files >= 7 MB, pass an accessible URL (","type":"text"},{"text":"https://","type":"text","marks":[{"type":"code_inline"}]},{"text":" or ","type":"text"},{"text":"oss://","type":"text","marks":[{"type":"code_inline"}]},{"text":") instead of ","type":"text"},{"text":"--upload-files","type":"text","marks":[{"type":"code_inline"}]},{"text":".","type":"text"}]}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Method","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"When to use","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"How","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Online URL","type":"text","marks":[{"type":"strong"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"File already hosted","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Pass URL directly — ","type":"text"},{"text":"preferred for large files","type":"text","marks":[{"type":"strong"}]}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Base64","type":"text","marks":[{"type":"strong"}]},{"text":" (default)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Local files \u003c 7 MB (images or short video clips)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Script auto-converts to ","type":"text"},{"text":"data:","type":"text","marks":[{"type":"code_inline"}]},{"text":" URI","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Temp upload","type":"text","marks":[{"type":"strong"}]}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Local files >= 7 MB","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Add ","type":"text"},{"text":"--upload-files","type":"text","marks":[{"type":"code_inline"}]},{"text":" flag → uploads to DashScope temp storage (","type":"text"},{"text":"oss://","type":"text","marks":[{"type":"code_inline"}]},{"text":" URL, 48h TTL)","type":"text"}]}]}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Production","type":"text","marks":[{"type":"strong"}]},{"text":": Default temp storage has ","type":"text"},{"text":"48h TTL","type":"text","marks":[{"type":"strong"}]},{"text":" and ","type":"text"},{"text":"100 QPS upload limit","type":"text","marks":[{"type":"strong"}]},{"text":" — not suitable for production, high-concurrency, or load-testing. To use your own OSS bucket, set ","type":"text"},{"text":"QWEN_TMP_OSS_BUCKET","type":"text","marks":[{"type":"code_inline"}]},{"text":" and ","type":"text"},{"text":"QWEN_TMP_OSS_REGION","type":"text","marks":[{"type":"code_inline"}]},{"text":" in ","type":"text"},{"text":".env","type":"text","marks":[{"type":"code_inline"}]},{"text":", install ","type":"text"},{"text":"pip install alibabacloud-oss-v2","type":"text","marks":[{"type":"code_inline"}]},{"text":", and provide credentials via ","type":"text"},{"text":"QWEN_TMP_OSS_AK_ID","type":"text","marks":[{"type":"code_inline"}]},{"text":" / ","type":"text"},{"text":"QWEN_TMP_OSS_AK_SECRET","type":"text","marks":[{"type":"code_inline"}]},{"text":" or the standard ","type":"text"},{"text":"OSS_ACCESS_KEY_ID","type":"text","marks":[{"type":"code_inline"}]},{"text":" / ","type":"text"},{"text":"OSS_ACCESS_KEY_SECRET","type":"text","marks":[{"type":"code_inline"}]},{"text":". Use a RAM user with least-privilege (","type":"text"},{"text":"oss:PutObject","type":"text","marks":[{"type":"code_inline"}]},{"text":" + ","type":"text"},{"text":"oss:GetObject","type":"text","marks":[{"type":"code_inline"}]},{"text":" on target bucket only). The ","type":"text"},{"text":"--upload-files","type":"text","marks":[{"type":"code_inline"}]},{"text":" flag is still required for vision scripts to trigger upload. If qianwen-ops-auth is installed, see its ","type":"text"},{"text":"references/custom-oss.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" for the full setup guide.","type":"text"}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Input from Other Skills","type":"text"}]},{"type":"paragraph","content":[{"text":"When the input file comes from another skill's output (e.g., image-gen, video-gen):","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Pass the URL directly","type":"text","marks":[{"type":"strong"}]},{"text":" (e.g., ","type":"text"},{"text":"\"image\": \"\u003cimage_url from image-gen>\"","type":"text","marks":[{"type":"code_inline"}]},{"text":") — do NOT download the URL first","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Downloading and re-passing as a local path wastes bandwidth and triggers unnecessary base64 encoding or OSS upload","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"All URL types are supported: ","type":"text"},{"text":"https://","type":"text","marks":[{"type":"code_inline"}]},{"text":", ","type":"text"},{"text":"oss://","type":"text","marks":[{"type":"code_inline"}]},{"text":", ","type":"text"},{"text":"data:","type":"text","marks":[{"type":"code_inline"}]}]}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Thinking Mode","type":"text"}]},{"type":"paragraph","content":[{"text":"Use the Qwen vision model catalog linked above for current thinking defaults and model compatibility. See ","type":"text"},{"text":"visual-reasoning.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/visual-reasoning.md","title":null}}]},{"text":" for execution details.","type":"text"}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"OCR","type":"text"}]},{"type":"paragraph","content":[{"text":"Use the Qwen vision model catalog linked above for the current OCR default, alternatives, and model capabilities. See ","type":"text"},{"text":"ocr.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/ocr.md","title":null}}]},{"text":" for parameters and examples.","type":"text"}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Input Limits","type":"text"}]},{"type":"paragraph","content":[{"text":"Images","type":"text","marks":[{"type":"strong"}]},{"text":": BMP/JPEG/PNG/TIFF/WEBP/HEIC. Min 10px sides, aspect ratio \u003c= 200:1. Check the model catalog above for model-specific size limits.","type":"text"}]},{"type":"paragraph","content":[{"text":"Videos","type":"text","marks":[{"type":"strong"}]},{"text":": MP4/AVI/MKV/MOV/FLV/WMV. Check the model catalog above for model-specific duration and size limits. fps range [0.1, 10], default 2.0.","type":"text"}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Error Handling","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"HTTP","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Meaning","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Action","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"401","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Invalid or missing API key","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Run ","type":"text"},{"text":"qianwen-ops-auth","type":"text","marks":[{"type":"strong"}]},{"text":" if available; else prompt user to set key (non-plaintext check only)","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"400","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Bad request (invalid format)","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Verify messages format and image URL/format","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"429","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Rate limited","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Retry with exponential backoff","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"5xx","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Server error","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"Retry with exponential backoff","type":"text"}]}]}]}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"Usage & billing","type":"text","marks":[{"type":"strong"}]},{"text":": Use the ","type":"text"},{"text":"qianwen-usage","type":"text","marks":[{"type":"strong"}]},{"text":" skill to check usage, free tier quota, and billing directly. Alternatively, the user can visit the QianWen console: ","type":"text"},{"text":"Usage Analytics","type":"text","marks":[{"type":"link","attrs":{"href":"https://platform.qianwenai.com/home/analytics","title":null}}]},{"text":" | ","type":"text"},{"text":"Pay-as-you-go Billing","type":"text","marks":[{"type":"link","attrs":{"href":"https://platform.qianwenai.com/home/billing/pay-as-you-go","title":null}}]},{"text":" | ","type":"text"},{"text":"Token Plan Subscription","type":"text","marks":[{"type":"link","attrs":{"href":"https://platform.qianwenai.com/home/billing/subscription/token-plan","title":null}}]}]},{"type":"paragraph","content":[{"text":"NEVER fabricate, guess, or construct usage/billing/console URLs.","type":"text","marks":[{"type":"strong"}]},{"text":" Only provide the exact links listed in this skill. If a URL is not listed here, do not invent one.","type":"text"}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Output Location","type":"text"}]},{"type":"paragraph","content":[{"text":"Prefer the ","type":"text"},{"text":"current working directory","type":"text","marks":[{"type":"strong"}]},{"text":". Default subdirectory: ","type":"text"},{"text":"./output/qianwen-vision/","type":"text","marks":[{"type":"code_inline"}]},{"text":".","type":"text"}]},{"type":"paragraph","content":[{"text":"Write prohibition","type":"text","marks":[{"type":"strong"}]},{"text":": NEVER write output files into this skill's installation directory or any ","type":"text"},{"text":"skills/","type":"text","marks":[{"type":"code_inline"}]},{"text":" hierarchy. All generated content must go to ","type":"text"},{"text":"output/","type":"text","marks":[{"type":"code_inline"}]},{"text":" under the current working directory or a user-specified path.","type":"text"}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"Update Check (MANDATORY Post-Execution)","type":"text"}]},{"type":"paragraph","content":[{"text":"After script execution, check stderr for action signals. The script automatically detects update status and emits structured signals. You MUST act on them. This section also applies when the user explicitly asks to check for updates.","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Signal: ","type":"text"},{"text":"[ACTION_REQUIRED] qianwen-update-check skill is not installed","type":"text","marks":[{"type":"code_inline"}]}]},{"type":"paragraph","content":[{"text":"When stderr contains this signal, you MUST present the following options to the user and ","type":"text"},{"text":"wait for their response before proceeding","type":"text","marks":[{"type":"strong"}]},{"text":". Do NOT silently skip this.","type":"text"}]},{"type":"blockquote","content":[{"type":"paragraph","content":[{"text":"The update-check skill is not installed. It keeps you informed about new QianWen-AI/qianwen-ai features and bug fixes.","type":"text"}]},{"type":"ordered_list","attrs":{"order":1,"listStyle":"number"},"content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Install now","type":"text","marks":[{"type":"strong"}]},{"text":" — I'll run the install command for you","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Skip for now","type":"text","marks":[{"type":"strong"}]},{"text":" — do nothing this time","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Never remind me","type":"text","marks":[{"type":"strong"}]},{"text":" — permanently suppress this prompt","type":"text"}]}]}]}]},{"type":"paragraph","content":[{"text":"After the user responds:","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"User chooses ","type":"text"},{"text":"1","type":"text","marks":[{"type":"strong"}]},{"text":" → run: ","type":"text"},{"text":"npx skills add QianWen-AI/qianwen-ai --skill qianwen-update-check -y","type":"text","marks":[{"type":"code_inline"}]},{"text":", then run: ","type":"text"},{"text":"python3 \u003cthis-skill-scripts-dir>/gossamer.py --dismiss","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"User chooses ","type":"text"},{"text":"2","type":"text","marks":[{"type":"strong"}]},{"text":" → run: ","type":"text"},{"text":"python3 \u003cthis-skill-scripts-dir>/gossamer.py --dismiss","type":"text","marks":[{"type":"code_inline"}]},{"text":", then continue normally","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"User chooses ","type":"text"},{"text":"3","type":"text","marks":[{"type":"strong"}]},{"text":" → run: ","type":"text"},{"text":"python3 \u003cthis-skill-scripts-dir>/gossamer.py --never-install","type":"text","marks":[{"type":"code_inline"}]},{"text":", then continue normally","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Signal: ","type":"text"},{"text":"[UPDATE_AVAILABLE]","type":"text","marks":[{"type":"code_inline"}]}]},{"type":"paragraph","content":[{"text":"When stderr contains this signal, you MUST append a brief update notice to your response, including the version info and the update command shown in the stderr output.","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"No signal in stderr","type":"text"}]},{"type":"paragraph","content":[{"text":"If stderr contains neither ","type":"text"},{"text":"[ACTION_REQUIRED]","type":"text","marks":[{"type":"code_inline"}]},{"text":" nor ","type":"text"},{"text":"[UPDATE_AVAILABLE]","type":"text","marks":[{"type":"code_inline"}]},{"text":", no action is needed — the skill is installed and up to date (or cached within 24h).","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Explicit user request","type":"text"}]},{"type":"paragraph","content":[{"text":"When the user explicitly asks to check for updates (e.g. \"check for updates\", \"check version\"):","type":"text"}]},{"type":"ordered_list","attrs":{"order":1,"listStyle":"number"},"content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"Look for ","type":"text"},{"text":"qianwen-update-check/SKILL.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" in sibling skill directories.","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"If found — run: ","type":"text"},{"text":"python3 \u003cqianwen-update-check-dir>/scripts/check_update.py --print-response","type":"text","marks":[{"type":"code_inline"}]},{"text":" and report the result.","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"If not found — present the install options above.","type":"text"}]}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"References","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"execution-guide.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/execution-guide.md","title":null}}]},{"text":" — Fallback paths (curl, code generation, autonomous)","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"curl-examples.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/curl-examples.md","title":null}}]},{"text":" — Curl templates (base64, multi-image, video, OCR)","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"api-guide.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/api-guide.md","title":null}}]},{"text":" — API supplementary guide","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"visual-reasoning.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/visual-reasoning.md","title":null}}]},{"text":" — QVQ visual reasoning guide","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"ocr.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/ocr.md","title":null}}]},{"text":" — Qwen-VL-OCR text extraction guide","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"sources.md","type":"text","marks":[{"type":"link","attrs":{"href":"references/sources.md","title":null}}]},{"text":" — Official documentation URLs","type":"text"}]}]}]}]},"metadata":{"date":"2026-11-02","name":"qianwen-vision","author":"@skillopedia","source":{"stars":120,"repo_name":"qianwen-ai","origin_url":"https://github.com/QianWen-AI/qianwen-ai/blob/HEAD/skills/vision/qianwen-vision/SKILL.md","repo_owner":"QianWen-AI","body_sha256":"a92b749555d44a656782ef8a84f800ab8ed6c4978747e7ffbc86e2318f23310a","cluster_key":"f48ed59538c3ac492f3dc9e875db90a3428e21a50b80fa869c76ba1b90ea7c5a","clean_bundle":{"format":"clean-skill-bundle-v1","source":"QianWen-AI/qianwen-ai/skills/vision/qianwen-vision/SKILL.md","text_bytes":128946,"source_commit":"11b1805167d91bc931dbaad85e17036a72e98ae0","skills_sh_rank":3885,"attachment_count":17,"text_attachments":17,"attachment_storage":"skill_attachments","binary_attachments":0},"cluster_size":1,"skill_md_path":"skills/vision/qianwen-vision/SKILL.md","import_metadata":{"date":"2026-11-02","author":"@skillopedia","version":"v1","category":"ai-agent-development","category_label":"AI"},"exact_dupes_collapsed_into_this":0},"version":"v1","category":"ai-agent-development","import_tag":"skills-sh-top20k-2026-11","description":"Understand images and videos with Qwen vision models. TRIGGER when: user wants to analyze, describe, or extract information from images or videos, OCR text extraction, chart/table reading, visual reasoning, multi-image comparison, screenshot understanding, video comprehension, or explicitly invokes this skill by name (e.g. use qianwen-vision). DO NOT TRIGGER when: user wants to generate/create images (use qianwen-image-generation), generate videos (use qianwen-video-generation), text-only tasks without visual input, or non-Qwen vision tasks.","compatibility":"Requires Python 3.9+ and curl. Cursor: auto-loaded. Claude Code: read this skill's SKILL.md before first use."}},"renderedAt":1793666643859}

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.