tryanswer

Naming Anything — Data skill for Claude Code

Data community

Agent Skill that names anything — apps, brands, shops, even people.

How to install Naming Anything

This entry records only its repository, not the path inside it, so there is no exact command to give. Open tryanswer/naming-anything and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Naming Anything does

Agent Skill that names anything — apps, brands, shops, even people. Four-stage pipeline · trademark-aware scoring · graded collision check (查重) · traditional 五行·八卦 layer. Works with any SKILL.md-compatible agent.

Alternatives in Data

Private GPT — Complete API layer for private AI applications on local models: RAG, skills, tools, MCP, text-to-sql, and more 57.5k ★
HTML Anything — ✨ The agentic HTML editor — your local AI agent writes the HTML, you ship it 8.5k ★
Osint Employees — Gather employee names + email patterns for password spray prep 4.5k ★

README

naming-anything

[![License: MIT](https://img.shields.io/badge/License-MIT-3fb950.svg)](LICENSE) [![Type: Agent Skill](https://img.shields.io/badge/type-Agent_Skill-8957e5.svg)](https://github.com/anthropics/skills) [![传统层：五行 · 三才五格 · 八卦](https://img.shields.io/badge/传统层-五行_·_三才五格_·_八卦-db6d28.svg)](#4-traditional-chinese-scoring-layer-五行--三才五格--八卦) [![GitHub stars](https://img.shields.io/github/stars/tryanswer/naming-anything.svg?style=flat&color=ff8c42)](https://github.com/tryanswer/naming-anything/stargazers)

An Agent Skill for naming anything — apps, products, companies, open-source projects, offline shops, events, pets, IPs, and people — through a professional four-stage pipeline, with bilingual (Chinese-English) output, trademark-aware scoring, a graded collision screen (查重), and an optional traditional Chinese metaphysics layer (**五行 · 三才五格 · 八卦**).

Packaged as an Agent Skill (the open `SKILL.md` convention): any agent that loads skills from a skills directory can use it, unmodified.

1. Overview

LLM-assisted naming has a characteristic failure mode: polished-but-interchangeable candidates pulled from thesaurus reflexes and lazy suffixes ("-ly", "-ify", "智慧XX", "XX管家") — names that could sit on any product in any category. **naming-anything** counters this with the working discipline of professional naming agencies, adapted for agent execution:

Method before inspiration. Candidates are generated inside mapped metaphor territories and across seven structured methods, never by free-associating on a category word.
Law before aesthetics. Every candidate is placed on the trademark Distinctiveness Spectrum (Generic → Descriptive → Suggestive → Arbitrary → Fanciful); the Suggestive sweet spot is the default target for consumer-facing objects.
Evidence before delivery. Finalists pass a four-query collision screen with graded verdicts and a repair playbook, and the delivery separates what was actually checked from what remains a human