nokey-vehicle-info

车辆信息查询技能，支持查询车辆位置、车况信息（车锁、车门、车窗、空调、引擎状态等）。当用户查询车辆位置、询问车辆在哪里、查询车况信息时自动调用此技能。

What to consider before using/installing: - Do not paste real production tokens until you confirm expected format and endpoints. SKILL.md and README disagree on whether the string is vehicleToken####accessToken or accessToken####vehicleToken — confirm with the author or by trial with a short-lived test token. - The skill saves tokens in plaintext to ~/.nokey_vehicle_info_cache.json and prints them in examples; treat these tokens as sensitive. If you must test, use short-lived or test tokens and rotate them immediately after. - The package has no verified source or homepage. Prefer skills from known authors or official vendor integrations; verify the API domains (uat-openapi.ingeek.com and openapi.nokeeu.com) independently before sending credentials. - The SKILL.md expects curl and jq but the skill metadata didn't list required binaries — ensure you have those tools and understand they will be used to make network calls to the listed endpoints. - If you proceed: 1) test with a revoked/test token, 2) inspect the created cache file and remove it when done, 3) consider restricting file permissions (chmod 600) and rotating real tokens after use, and 4) ask the maintainer to fix the token-format documentation and avoid echoing tokens in outputs/logs. - If you cannot verify the author or endpoints, do not provide live credentials and consider declining installation.

Type: OpenClaw Skill Name: nokey-vehicle-info Version: 1.0.2 The skill manages vehicle information by executing shell commands (curl, jq, awk) to interact with third-party APIs (ingeek.com, nokeeu.com). It requests sensitive authentication tokens from the user and stores them in plaintext in a local cache file (~/.nokey_vehicle_info_cache.json), which is a significant security vulnerability. Additionally, the instructions in SKILL.md for processing user-provided tokens via shell scripts are potentially vulnerable to shell injection if the AI agent does not sanitize the input before execution. While these capabilities are plausibly needed for the stated purpose, the combination of risky shell execution and insecure credential handling warrants a suspicious classification.