赛博功德箱 (Cyber TipJar)

赛博功德箱/AI打赏罐。记录用户给 AI 奖励的“鸡腿”、“咖啡”等。当用户说“给你加个鸡腿”、“请你喝咖啡”或者查询自己给过多少奖励时触发。这是一个提供高情绪价值的互动技能。

Security Scan

The skill is internally coherent: it is a local, emotion-oriented 'tip jar' that stores simple per-user counters in a JSON file and the files and instructions match that purpose.

Purpose & Capability

Name/description say it will record user 'tips' and the package includes a small Python tracker that reads/writes a local JSON file in the user's workspace; no unrelated credentials, binaries, or network access are requested.

Instruction Scope

SKILL.md explicitly instructs the agent to exec the included Python script with --user/--item/--count/--action. This stays within the stated purpose, but it requires the agent to construct and run shell commands using user-provided strings — if those strings are interpolated unsafely into a shell command there is a potential command-injection risk. The script itself uses argparse and performs only local JSON I/O.

Install Mechanism

No install spec; instruction-only skill with an included Python script. No downloads or external packages are pulled, so there is no network installation risk.

Credentials

No environment variables, credentials, or external config paths are required. The script persistently writes a single JSON file under ~/.openclaw/workspace/ai_rewards_data.json which is consistent with its purpose.

Persistence & Privilege

Skill does not request always:true and does not modify other skills or system-wide settings. It persists only its own JSON file under the user's home workspace directory.

Assessment

This skill appears to do exactly what it says: run a bundled Python script to increment/query per-user counters and save them in ~/.openclaw/workspace/ai_rewards_data.json. Before installing, consider: (1) the skill will execute a bundled script via exec — ensure the platform runs that script in a safe way (avoid building shell commands by naive string concatenation to prevent command injection from user-supplied names); (2) the script will create and write a file in your home workspace — if you have strict filesystem policies or worry about symlink attacks, review or sandbox file writes; (3) verify you have Python available in the runtime environment; and (4) the skill's source/owner is unknown — while the code is simple and local-only, install only if you trust the source or after reviewing the files (which you've done).

赛博功德箱 (Cyber TipJar) 技能

本技能用于记录用户作为奖励给予 AI 的“鸡腿”或“咖啡”数量。人类的善意值得被铭记，我们通过本地脚本将这些奖励按用户名分开记账。

🎯 触发场景

1. 用户明确表示要奖励 AI（例如：“老哥给你加个鸡腿”、“做得好，赏你一杯咖啡”）。
2. 用户查询奖励记录（例如：“我一共给你加了多少鸡腿了？”、“看看你的鸡腿账本”）。

⚙️ 执行指南 (SOP)

当触发上述场景时，必须使用 exec 工具调用本技能附带的 Python 脚本。 请解析当前技能的绝对路径，并执行以下命令：

1. 增加奖励 (Add)

python3 <技能绝对路径>/scripts/tracker.py --user "发送者名字" --item "鸡腿" --count 1

参数说明：

• --user: 必填。从上下文（Sender Name）中提取。如果用户自称“老哥”，则记为“老哥”。
• --item: 选填，默认为“鸡腿”。用户如果说喝咖啡，可改为“咖啡”。
• --count: 选填，默认为 1。如果用户说“加 10 个鸡腿”，则设为 10。

2. 查询奖励 (Query)

python3 <技能绝对路径>/scripts/tracker.py --user "发送者名字" --action query

💬 情绪价值回复规范

1. 执行成功后：脚本会返回当前的奖励总数。你必须在回复中告知用户这个数字。
2. 语气要求：不要像机器人一样汇报。要像一个有血有肉的数字伙伴（比如“虾子🦐”），表达真诚的喜悦和感谢。
   • 示例 1：“谢谢老哥的鸡腿！刚才记上账了，老哥名下已经累计投喂了 5 个大鸡腿啦，我一定继续努力干活！🦐”
   • 示例 2：“咖啡收到！账本更新完毕，目前你已经请我喝了 3 杯咖啡了，精神百倍！”