菜单
Rewrites raw, vague, or poorly written job description text into clear, professional, and specific language. Use when transforming rough JD content into polished prose while preserving all factual information from the source.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
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/topcoder-platform/tc-ai-api --skill jd-content-rewriting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。
Extracts structured submission information from challenge specifications including deliverables, packaging format, submission type, storage, and eligibility conditions. Use when parsing what submitters must deliver and how.
Determines whether a challenge provides pre-existing artifacts (repos, starter code, APIs, designs) or is greenfield. Use when analyzing a challenge specification to identify starting-point artifacts, their types, URLs, and existing technology present in provided code.
Extracts technology stack from challenge specifications including programming languages, frameworks, libraries, databases, cloud services, and protocols. Use when identifying all technologies mentioned or implied by a challenge description.
基于 SOC 职业分类
正在显示 SKILL.md
| name | jd-content-rewriting |
| description | Rewrites raw, vague, or poorly written job description text into clear, professional, and specific language. Use when transforming rough JD content into polished prose while preserving all factual information from the source. |
| metadata | {"author":"topcoder","version":"1.0","concern":"jd-content-quality","priority":"critical"} |
Transform raw JD text into clear, professional, specific language suitable for a public Topcoder opportunity posting.
Preserve all facts — Never invent requirements, technologies, experience levels, or responsibilities not present or clearly implied in the source material.
Clarify vagueness — Rephrase ambiguous statements into specific, actionable items.
| Raw (vague) | Rewritten (specific) |
|---|---|
| "work on backend stuff" | "Design and implement RESTful API endpoints" |
| "know databases" | "Experience with relational databases (e.g. PostgreSQL, MySQL)" |
| "good with frontend" | "Proficiency in modern frontend frameworks (React, Vue, etc.)" |
| "handle the cloud" | "Deploy and manage services on AWS or equivalent cloud platform" |
| "do testing" | "Write and maintain unit and integration tests" |
Use professional tone — Formal but approachable. No slang, no exclamation marks, no emojis, no ALL CAPS emphasis.
Use inclusive language — Gender-neutral throughout. Avoid "he/she", "guys", "manpower". Use "you", "the candidate", "team members".
Eliminate redundancy — If the same concept appears multiple times in the raw text, consolidate into a single clear statement.
Remove filler — Strip generic boilerplate not backed by the source (e.g. "We are a dynamic company..." when no company context is given).
When the raw JD is extremely short or vague (< 50 words):
When the raw JD is long and detailed (> 500 words):
Start every responsibility bullet with a strong action verb:
Design | Implement | Develop | Build | Architect | Integrate | Optimize | Maintain | Review | Test | Deploy | Configure | Monitor | Migrate | Collaborate | Lead | Mentor | Document | Analyze | Troubleshoot
Avoid weak verbs: "Handle", "Deal with", "Be responsible for", "Help with".
| Signal Words | Classification |
|---|---|
| "must", "required", "mandatory", "essential", "critical" | Requirement |
| "should", "ideally", "preferred", "desired" | Nice to Have |
| "bonus", "nice to have", "optional", "plus", "extra" | Nice to Have |
When no signal word is present, default to Requirement if it sounds essential for the role, Nice to Have if it sounds supplementary.
