Agent skill
metadata-naming
Define, apply, or review a reusable metadata-based filename standard for files, folders, inventories, archives, and generated catalogs. Use when the user wants to extract a naming convention into a standard, normalize filenames, choose between human-readable and machine-friendly naming, or create naming rules for repositories, skills, software inventories, snapshots, and long-lived records.
Install this agent skill to your Project
npx add-skill https://github.com/luw2007/skills/tree/main/metadata-naming
SKILL.md
Metadata Naming
Turn loose filename habits into a stable naming standard that is easy to read, sort, parse, and reuse.
Use This Standard
Follow this split by file purpose:
- Use stable identity naming for long-lived entries that are updated in place.
- Use timestamp naming for snapshots, archives, exports, and reports.
- Prefer ASCII, no spaces, and fixed separator rules when filenames will be processed by scripts or moved across systems.
Core Metadata Blocks
Use these blocks in a fixed order when a filename needs richer metadata:
- time
- prefix
- title
- version
- tags
- source_or_author
- note
Not every block is required. Keep only the fields that improve retrieval or automation.
Two Modes
Relaxed mode
Use for human-managed documents where readability matters more than strict parsing.
Conventions:
- Chinese or mixed-language titles are allowed.
- Spaces are allowed when the surrounding system tolerates them.
- Typical visual markers:
- time:
(2026-03-11)or(20260311-093500) - version:
(v1.2.0) - tags:
#tag - source or author:
@name - note:
¬e
- time:
Example:
(2026-03-11)Favorites Curator(v0.1.0)#skill#favorites@workspace&initial publish
Strict mode
Use by default for standards, skills, inventories, generated files, syncable folders, and anything a script will read.
Conventions:
- ASCII only
- no spaces
- top-level separator:
__ - intra-block separator:
-or. - lowercase slug-style titles unless there is a strong reason not to
General template:
YYYYMMDD[-HHMMSS]__prefix__title__version__tags__source__note.ext
Examples:
20260311__skill__favorites-curator__v0.1.0__favorites.catalog__workspace.md
20260311__repo__openclaw-backup-tool__v0.1.0__backup.tool__github.md
20260311__app__codex__v0.112.0__cli.ai__brew.md
Default Rule For Long-Lived Entries
For catalogs, inventories, and canonical records, prefer stable filenames over timestamped filenames.
Use this template:
<data_type>__<source_name>__<slug>.md
Examples:
skill__workspace__favorites-curator.md
repo__github__openclaw-backup-tool.md
app__brew__codex.md
Use this rule when the content is refreshed in place and the filename should not drift over time.
Default Rule For Snapshots And Reports
Use timestamp-first filenames for time-series artifacts.
Templates:
YYYYMMDD__report__topic.md
YYYYMMDD-HHMMSS__snapshot__topic.json
Examples:
20260311__report__favorites-digest.md
20260311-095914__snapshot__favorites.json
Block Rules
Time
- Use
YYYYMMDDfor day-level tracking. - Use
YYYYMMDD-HHMMSSfor run-level uniqueness. - Put time first when sort order matters.
Prefix
- Use short taxonomy values such as
repo,skill,app,doc,snapshot,report. - Keep the vocabulary stable once chosen.
Title
- Make this the main identity.
- Prefer short, stable, searchable slugs.
- Move extra description into tags or notes.
Version
- Use semver when available:
v0.1.0,v2.3.4. - Omit the block when versioning is irrelevant.
Tags
- Use compact, low-noise tags.
- Join multiple tags with
.or-inside the same block.
Source Or Author
- Use the source system, publisher, owner, or author when it improves retrieval.
- Examples:
github,brew,workspace,openclaw,vendor-name.
Note
- Keep it short.
- Do not put long prose into filenames.
- Use only when the note materially changes retrieval value.
Standard Principle
Use fixed-order metadata blocks. Use stable identity filenames for long-lived entries and timestamped filenames for snapshots. Default to ASCII, no spaces, __ between blocks, and - or . inside blocks so names stay sortable, parseable, and portable.
References
Read references/standard.md when you need the normalized standard, examples, and decision rules in a reference-friendly format.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
rtk-rewrite
RTK rewrite plugin for OpenClaw. Intercepts exec tool calls and delegates rewrites to rtk rewrite to reduce token usage while preserving command intent.
octocode
AI-powered code indexer with semantic search, knowledge graphs (GraphRAG), and persistent memory system. Use when you need to (1) perform semantic code searches across large codebases, (2) analyze file relationships and dependencies through GraphRAG, (3) store and retrieve code insights with memory system. This skill uses Bash to call octocode CLI directly with automatic index management.
favorites-curator
Build and maintain a local favorites catalog from installed repositories, apps, skills, extensions, and hooks. Use when the user asks to create, refresh, curate, scan, or summarize a favorites collection, software inventory, skill inventory, or daily digest of useful local resources.
langextract-search
集成智谱搜索、DuckDuckGo 搜索和多模型结构化提取的完整工作流。
skill-selector
通用 skill 选择器。通过 ai-skills CLI 下载候选 skill 源码到临时目录,基于 SKILL.md + 代码的深度分析构建决策树,引导用户在 1-3 轮交互内选定最合适的 skill。当用户面对多个功能相近的 skill 不知选哪个时触发。
collaborating-with-coco
Delegates coding tasks to Coco CLI (Codebase Copilot) for prototyping, debugging, and code review. Use when needing algorithm implementation, bug analysis, or code quality feedback. Supports multi-turn sessions via SESSION_ID.
Didn't find tool you were looking for?