← 返回 Skills 市场
Feather
作者
bytesagain1
· GitHub ↗
· v1.0.0
· MIT-0
136
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install feather
功能描述
Apache Feather/Arrow IPC format reference. V1 vs V2 format differences, pyarrow.feather read/write with compression, R arrow package integration, Arrow type...
使用说明 (SKILL.md)
Feather
Apache Feather/Arrow IPC format reference — fast columnar DataFrame I/O.
Commands
| Command | Description |
|---|---|
intro |
Feather overview, V1 vs V2, key properties |
python |
pyarrow.feather read/write, benchmarks |
r-lang |
R arrow package, Python↔R interop |
schema |
Arrow types, nested types, metadata |
vs-parquet |
Speed vs compression tradeoffs |
pandas |
pd.read_feather, caching patterns |
compression |
LZ4/ZSTD options, when to use each |
best-practices |
Pipeline patterns, file size guidelines |
安全使用建议
This skill is a read-only reference and appears safe: it only contains documentation, code examples, and a shell script that prints help text. It does not request credentials or perform network downloads. If you expect the agent to actually run the example commands (pip install, R install.packages, or pip/OS-level commands), be aware those will execute on the host/environment where the agent runs — the skill itself will not install dependencies. Also note a truncated section in the SKILL.md (trailing ellipsis) — functional behavior is documentation-only, but if you need executable tooling, ensure required libraries are installed from trusted package sources before running examples.
功能分析
Type: OpenClaw Skill
Name: feather
Version: 1.0.0
The 'feather' skill is a documentation and reference tool for the Apache Feather/Arrow IPC format. The files (SKILL.md and scripts/script.sh) contain only informational content, code snippets, and benchmarks related to data processing with Python and R. There is no executable logic beyond printing text to the console, and no evidence of data exfiltration, prompt injection, or malicious intent.
能力评估
Purpose & Capability
Name, description, SKILL.md and included script all focus on Feather/Arrow format, pyarrow and R arrow usage, benchmarks and best practices. There are no unexpected credentials, binaries, or unrelated capabilities requested.
Instruction Scope
The runtime instructions and the bundled script only present documentation, usage examples, and CLI-like help text. They do not instruct the agent to read arbitrary system files, access environment variables, or transmit data to external endpoints.
Install Mechanism
No install spec is present (instruction-only). The included shell script is documentation-only and does not download or execute remote code. This low-risk distribution method is proportionate for a reference skill.
Credentials
The skill requires no environment variables, no credentials, and references only typical library installs (e.g., pip install pyarrow, install.packages('arrow')) that are appropriate for the documented examples.
Persistence & Privilege
always is false and the skill does not request elevated or persistent privileges, nor does it modify other skills or system-wide configurations. Autonomous invocation is allowed but consistent with normal skill behavior.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install feather - 安装完成后，直接呼叫该 Skill 的名称或使用
/feather触发 - 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
v1.0.0
publish v1.0.0
元数据
常见问题
Feather 是什么？
Apache Feather/Arrow IPC format reference. V1 vs V2 format differences, pyarrow.feather read/write with compression, R arrow package integration, Arrow type... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 136 次。
如何安装 Feather？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install feather」即可一键安装，无需额外配置。
Feather 是免费的吗？
是的，Feather 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Feather 支持哪些平台？
Feather 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 Feather？
由 bytesagain1（@bytesagain1）开发并维护，当前版本 v1.0.0。
推荐 Skills