file_reader
Read and summarize text-based file types only. Prefer read_file for text formats; use execute_shell_command for type detection when needed. PDF/Offi…
它会碰到什么
这一栏是扫描器报的事实,不是结论。命中多不等于有毒(安全工具、规则库、示例脚本本来就会包含危险写法),命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系,需要你自己看。
技能内容
File Reader Toolbox
Use this skill when the user asks to read or summarize local text-based files. PDFs, Office documents, images, audio, and video are out of scope for this skill and should be handled by their dedicated skills/tools.
Quick Type Check
Use a type probe before reading:
file -b --mime-type "/path/to/file"
If the file is large, avoid dumping the whole content; extract a small, relevant portion and summarize.
Text-Based Files (use read_file)
Preferred for: .txt, .md, .json, .yaml/.yml, .csv/.tsv, .log, .sql, ini, toml, py, js, html, xml source code.
Steps:
- Use
read_fileto fetch content. - Summarize key sections or show the relevant slice requested by the user.
- For JSON/YAML, list top-level keys and important fields.
- For CSV/TSV, show header + first few rows, then summarize columns.
Large Logs
If the file is huge, use a tail window:
tail -n 200 "/path/to/file.log"
Summarize the last errors/warnings and notable patterns.
Out of Scope
Do not handle the following in this skill (they are covered by other skills):
- Office (docx/xlsx/pptx)
- Images
- Audio/Video
Safety and Behavior
- Never execute untrusted files.
- Prefer reading the smallest portion necessary.
- If a tool is missing, explain the limitation and ask the user for an alternate format.
想直接用这个技能?
本站把开放许可(MIT / Apache 等)的技能按仓库打包整理到网盘,点一下转存到你自己的网盘,不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接,不打包。
它属于哪个仓库
src/qwenpaw/agents/skills/file_reader-en/SKILL.md同一个仓库里的其他技能
同名技能的其他版本
有 2 个不同仓库或目录里都有叫 file_reader 的技能。它们内容并不相同,别混用:
- agentscope-ai/QwenPaw — 仅用于读取和总结文本类文件。文本格式优先使用 read_file;需要做类型探测时使用 execute_shell_command。PDF、Office、图片和压缩包由其他技能处理