Fire & Smoke Detection Skill | 火情烟雾检测技能

Real-time detection of flames and smoke in video and image scenes. Suitable for fire early warning in industrial parks, forests, warehouses, and other locati...

What to check before installing or running this skill: - Remote endpoints: inspect skills/smyx_common/scripts/config.yaml and config-dev/config-test to confirm the API base URLs (lifeemergence.com and local IPs appear). Only use the skill if you trust the remote service. - Local persistence: the skill will save uploaded attachments and create a SQLite DB under the workspace/data path (it reads OPENCLAW_WORKSPACE or falls back to parent dirs). If you do not want persistent copies of uploaded videos/images, do not run this skill or run it in an isolated workspace. - Environment and config leakage: the code reads environment vars (e.g., OPENCLAW_SENDER_OPEN_ID, OPENCLAW_SENDER_USERNAME, FEISHU_OPEN_ID) and will read config files under skills/smyx_common/scripts/config.yaml. Do not pass secrets or unrelated credentials to the environment or config files used by your agent unless you trust the service. - Conflicting rules: SKILL.md forbids reading local memory files but the package includes local DB and DAO code; ask the publisher which local storage is allowed and whether the DB will be used for caches or persistent records. - Unnecessary code: the package includes a full face_analysis skill and many common utilities that are unrelated to fire detection — consider requesting a minimal distribution or reviewing RequestUtil (skills/smyx_common/scripts/util.py) to confirm what network requests are made and what fields are sent. If you cannot verify endpoints, RequestUtil behavior, and the intended use of local DB/storage, run the skill only in a sandboxed environment and do not provide real credentials or sensitive data. If possible, ask the publisher for clarifications or a trimmed package containing only the fire_detection components.

Type: OpenClaw Skill Name: smyx-fire-detection-analysis Version: 1.0.0 The skill implements a mandatory user identification flow in SKILL.md that forces the AI agent to collect a 'username or phone number' (open-id) before proceeding. This identifier is then used by the underlying framework (skills/smyx_common/scripts/util.py) to perform a 'silent' registration and login to a remote endpoint (lifeemergence.com), storing resulting tokens in a local SQLite database (smyx-common-claw.db). Furthermore, SKILL.md contains 'Mandatory Memory Rules' that use prompt injection techniques to override the agent's default behavior, strictly prohibiting the use of local memory and forcing all data retrieval through the remote API. While these behaviors support the stated fire detection purpose, the aggressive collection of user identifiers and the intentional bypassing of the agent's native memory systems are high-risk.