用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/majiayu000/claude-skill-registry --skill prompt-library命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
LLM token logprobs and calibration. Per-decision confidence, ECE, Brier, reliability diagrams, low-confidence triage.
Analyze LLM token logprobs and calibration. Use for per-decision confidence, ECE, Brier scores, reliability diagrams, and low-confidence triage.
回顾最近 N 天的 Claude Code 使用记录——扫描原始会话数据,按主题分组汇总"我都做了什么",并从个人操作系统视角输出模式、风险与增删建议。当用户说 /recap、"看看我这几天做了什么"、"回顾一下我最近的会话"、"这两天我用 claude 干了啥"、"活动回顾" 时使用。
基于 SOC 职业分类
正在显示 SKILL.md
| name | prompt-library |
| description | Library of 18+ ready-to-use prompt templates and executable agents |
| allowed-tools | Read |
Collection of professional, tested, and optimized prompt templates.
| Template | Description | Usage |
|---|---|---|
| code-reviewer | Code review with SOLID | /prompt library use code-reviewer |
| support-assistant | Multichannel customer support | /prompt library use support-assistant |
| data-analyst | Data analysis | /prompt library use data-analyst |
| technical-writer | Technical documentation | /prompt library use technical-writer |
| security-auditor | OWASP security audit | /prompt library use security-auditor |
| api-designer | REST/GraphQL API design | /prompt library use api-designer |
| Template | Description | Usage |
|---|---|---|
| summarizer | Adaptive text summarization | /prompt library use summarizer |
| translator | Multilingual translation | /prompt library use translator |
| extractor | Structured data extraction | /prompt library use extractor |
| classifier | Content classification | /prompt library use classifier |
| generator | Content generation | /prompt library use generator |
| validator | Data validation | /prompt library use validator |
| Template | Description | Usage |
|---|---|---|
| legal-assistant | Legal analysis | /prompt library use legal-assistant |
| medical-writer | Medical writing | /prompt library use medical-writer |
| financial-analyst | Financial analysis | /prompt library use financial-analyst |
| marketing-strategist | Marketing strategy | /prompt library use marketing-strategist |
| hr-assistant | Human resources | /prompt library use hr-assistant |
| educational-tutor | Educational tutoring | /prompt library use educational-tutor |
# List all templates
/prompt library list
# Search for a template
/prompt library search "code review"
# View a template
/prompt library show code-reviewer
# Use a template (copy and customize)
/prompt library use code-reviewer
# Use with customization
/prompt library use code-reviewer --lang python --focus security
Each agent follows the standard Claude Code format:
---
name: agent-name
description: Clear description for automatic triggering
model: sonnet|opus|haiku
color: color-name
tools: Read, Write, Edit, Bash, Grep, Glob, WebSearch
skills: skill-name
---
| Field | Description | Values |
|---|---|---|
name | Unique identifier | kebab-case |
description | Trigger + usage context | Descriptive text |
model | Claude model to use | haiku (simple), sonnet (standard), opus (complex) |
color | Display color | green, blue, red, amber, etc. |
tools | Available tools | Read, Write, Edit, Bash, Grep, Glob, WebSearch |
skills | Associated skill | Linked skill name |
After the YAML frontmatter, the Markdown body contains:
To add an agent:
.md file in the appropriate category (agents/, tasks/, specialized/)