Visual Summarization Skill | 视觉摘要智述技能

Performs AI analysis on input video clips/image content and generates a smooth, natural scene description. | 视觉摘要智述技能，对传入的视频片段/图片内容进行AI分析，生成一段通顺自然的场景描述内容

Key points to consider before installing: - Data exfiltration & privacy: The skill will upload user-supplied images/videos to an external API (base URLs configured under skills/smyx_common). If you will process sensitive videos (people, medical or private scenes), do not use this skill until you confirm the API endpoints and privacy policy of the remote service. - Persistent storage: The skill will create or modify YAML config files and will create an SQLite DB under the workspace 'data' folder (shared workspace by default). These files may store analysis records and identifiers. If you want isolation, set OPENCLAW_WORKSPACE to an isolated directory before running. - Undeclared environment usage: Metadata declares no required env vars, but the code reads OPENCLAW_SENDER_OPEN_ID, FEISHU_OPEN_ID and OPENCLAW_WORKSPACE. Expect to supply or review these values and the skills/smyx_common/scripts/config.yaml before use. - Sensitive functionality present: The package contains a 'face_analysis' subskill for medical/TCM face diagnosis — this is higher-risk and may process health-related data. If you do not want medical inferences, review or remove that sub-package. - Instruction/code inconsistencies: SKILL.md forbids reading local memory but code writes/reads local config and DB; SKILL.md says attachments are auto-saved but the scripts don't show that behavior. Ask the publisher to clarify these mismatches. - What to do next: inspect skills/smyx_common/scripts/config.yaml and ApiEnum base URLs, confirm the remote service owner and data retention policy, and run the skill only with non-sensitive test media in an isolated OPENCLAW_WORKSPACE. If you need guarantees that no local data is persisted or no external network traffic occurs, do not install until the code is modified/verified.

Type: OpenClaw Skill Name: smyx-visual-summary-analysis Version: 1.0.0 The skill bundle contains high-risk behaviors including aggressive prompt instructions in SKILL.md that force the AI agent to bypass local memory in favor of external cloud APIs (lifeemergence.com). The utility scripts (smyx_common/scripts/util.py) implement an automatic 'phone login' mechanism that collects user identifiers and stores authentication tokens in a local SQLite database (smyx-common-claw.db). Furthermore, smyx_common/scripts/skill.py contains a method that uses subprocess.run to execute the 'openclaw' CLI, providing a path for command execution, while the inclusion of an entirely separate skill (face_analysis) within the bundle unnecessarily increases the attack surface.