← 返回 Skills 市场
udiedrichsen

Streaming Buddy

作者 udiedrichsen · GitHub ↗ · v2.0.0
cross-platform ⚠ suspicious
2299
总下载
1
收藏
3
当前安装
1
版本数
在 OpenClaw 中安装
/install streaming-buddy
功能描述
Personal streaming assistant with learning preferences. Tracks what you're watching, learns your taste, and suggests what to watch next based on your services, mood, and preferences. Use when asked about movies, TV shows, streaming services, what to watch, or tracking viewing progress. Triggers: /stream, 'what should I watch', 'recommend something', mentioning Netflix/Prime/Disney+/Apple TV+, asking about series/seasons/episodes, mood-based requests like 'something exciting'.
安全使用建议
Things to check before installing: 1) API key handling: The skill's metadata claims TMDB_API_KEY is an environment variable, but the script reads the key from $WORKSPACE/memory/streaming-buddy/config.json. Decide which you prefer. If you want to use an environment variable, inspect/modify handler.sh to accept ${TMDB_API_KEY} from the environment (or store the key locally if you prefer not to expose it as a process env). 2) Undeclared utilities: The script uses md5sum and stat and relies on specific stat flags; ensure these are available on your platform (or adapt the script). The declared required binaries list only jq and curl — consider adding md5sum and verifying stat behavior. 3) Workspace permissions: The skill writes persistent data under the provided workspace path. Make sure that workspace is a directory you trust and that sensitive credentials are stored where you expect. The config.json file will contain your API key in cleartext if you follow SKILL.md instructions. 4) Network traffic: The handler performs outbound HTTPS calls to api.themoviedb.org (expected). The references mention JustWatch's GraphQL endpoint only as an alternative; the shipped handler does not call JustWatch directly. If you see any modifications to the script that call other endpoints, review them carefully. 5) Review the full handler script: We reviewed the visible portions and found no obfuscated code or hidden endpoints, but the script is sizable and truncated in the package listing — before trusting it, open and read the entire scripts/handler.sh to confirm its behavior matches expectations (especially preference updates, any telemetry, or unexpected external requests). If these issues (env-vs-config mismatch, undeclared md5sum/stat dependency) are acceptable and you host the workspace in a place you control, the skill's behavior is coherent with its purpose. If not, request or make the small fixes above before enabling it.
功能分析
Type: OpenClaw Skill Name: streaming-buddy Version: 2.0.0 The skill bundle is benign. The `handler.sh` script uses `curl` to interact with the legitimate TMDB API (`api.themoviedb.org`) for its stated purpose of providing streaming recommendations. All file operations are confined to the designated `$WORKSPACE/memory/streaming-buddy` directory, and there is no evidence of data exfiltration, malicious execution (e.g., `eval` with untrusted input, `curl|bash`), persistence mechanisms, or prompt injection attempts in `SKILL.md`.
能力评估
Purpose & Capability
Name/description match the implementation: the skill queries TMDB, tracks local watch data, and makes recommendations. Requiring curl and jq is appropriate for a shell-based TMDB client. The declared TMDB_API_KEY requirement is reasonable for the stated purpose — however, the metadata claims an environment variable while runtime instructions and the script expect the key in a workspace config file, which is an inconsistency.
Instruction Scope
Runtime instructions and the handler focus on TMDB queries and local storage under $WORKSPACE/memory/streaming-buddy/. The SKILL.md explicitly instructs the user to store the API key in memory/streaming-buddy/config.json and to use workspace paths for all data. The handler reads/writes only those workspace files and calls TMDB endpoints. There are no instructions to read unrelated system files or to send data to third-party endpoints other than TMDB (the references mention a JustWatch GraphQL endpoint only as an alternative, but the handler uses TMDB). The main scope issue is the mismatch between declared env-var requirement and the actual config-file usage.
Install Mechanism
No install spec (instruction-only + a script). Nothing is downloaded or installed by an install step; the only runtime artifacts are files under the provided workspace directory.
Credentials
The skill only needs a TMDB API key for its core functionality, which is proportionate. However, registry/metadata mark TMDB_API_KEY as a required environment variable while the handler reads the key from memory/streaming-buddy/config.json (not from the environment). This mismatch could lead to confusion and accidental misconfiguration. Also the script uses md5sum and stat (and other standard utilities) but only lists jq and curl as required binaries — md5sum/stat may not exist or behave the same across all platforms and were not declared as required.
Persistence & Privilege
always:false and the script confines data to the specified workspace directory. It does not request system-wide configuration changes or other skills' credentials. Autonomous invocation is allowed (platform default) but not combined here with broad credentials or global persistence.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install streaming-buddy
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /streaming-buddy 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v2.0.0
v2.0: Learning system, mood-based recommendations, match explanations
元数据
Slug streaming-buddy
版本 2.0.0
许可证
累计安装 4
当前安装数 3
历史版本数 1
常见问题

Streaming Buddy 是什么?

Personal streaming assistant with learning preferences. Tracks what you're watching, learns your taste, and suggests what to watch next based on your services, mood, and preferences. Use when asked about movies, TV shows, streaming services, what to watch, or tracking viewing progress. Triggers: /stream, 'what should I watch', 'recommend something', mentioning Netflix/Prime/Disney+/Apple TV+, asking about series/seasons/episodes, mood-based requests like 'something exciting'. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 2299 次。

如何安装 Streaming Buddy?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install streaming-buddy」即可一键安装,无需额外配置。

Streaming Buddy 是免费的吗?

是的,Streaming Buddy 完全免费(开源免费),可自由下载、安装和使用。

Streaming Buddy 支持哪些平台?

Streaming Buddy 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Streaming Buddy?

由 udiedrichsen(@udiedrichsen)开发并维护,当前版本 v2.0.0。

💬 留言讨论