upwork-proposal

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Upwork Proposal Skill

Upwork提案撰写技能

Generate proposals that beat the AI flood and actually get responses.

生成能够突破AI提案泛滥困境、真正获得回复的优质提案。

The Problem We're Solving

我们解决的问题

Clients receive 50+ proposals that all sound the same: robotic, vague, and stuffed with "Dear Hiring Manager." They can smell generic AI instantly. This skill creates proposals that feel human, specific, and worth responding to.

客户会收到50份以上千篇一律的提案：机械、模糊，满是“尊敬的招聘经理”这类套话。他们一眼就能识破通用AI生成的内容。本技能打造的提案更具人情味、针对性，值得客户回复。

Key Stats (2025-2026)

关键数据（2025-2026）

Metric	Value
Average proposals per job	20-50
Reply rate	8-30%
Win rate	~5% (1 in 20)
AI+Human hybrid win rate	48% higher than pure AI or pure manual
Optimal proposal length	150-250 words
First-hour applications	48% higher response rate

指标	数值
单职位平均提案数	20-50
回复率	8-30%
中标率	~5%（每20份中1份）
AI+人工混合提案中标率	比纯AI或纯人工提案高48%
最优提案长度	150-250词
发布首小时内提交的提案	回复率高48%

Workflow

工作流程

User pastes job → Extract requirements → Match to user skills → Generate proposal + shorter version

用户粘贴职位描述 → 提取需求 → 匹配用户技能 → 生成完整提案+精简版

Required Input

必填输入

Input	Description
job_description	The Upwork job posting (user pastes it)

输入项	说明
job_description	Upwork职位发布内容（用户粘贴）

Optional Inputs (Ask if relevant)

可选输入（按需询问）

Input	When to Ask
relevant_experience	If job matches a specific past project
rate/pricing	If user wants to include specific pricing
availability	If job mentions urgency or timeline
special_angle	If user has unique insight into the problem

输入项	询问时机
relevant_experience	当职位与用户某段过往项目经历匹配时
rate/pricing	当用户希望在提案中加入具体定价时
availability	当职位提及紧急性或时间线时
special_angle	当用户对项目问题有独特见解时

The 7-Part Formula (24%+ Response Rate)

七段式提案公式（回复率提升24%+）

See

references/proposal-structure.md

for details. Quick reference:

Part	Length	Purpose
Hook	25 words	Client name + their specific problem
Relevance	50-75 words	Similar experience with results
Solution	50-75 words	Bullet-point deliverables
Proof	30-50 words	1-2 past projects with metrics
Timeline	15-25 words	Specific dates, not vague
Investment	10-15 words	Clear pricing if appropriate
CTA	15-20 words	One specific next step

Total: 195-285 words (under 300 is ideal)

详情请见

references/proposal-structure.md

。快速参考：

部分	长度	目的
钩子	25词	提及客户姓名+他们的具体问题
相关性	50-75词	阐述类似经验及成果
解决方案	50-75词	分点列出交付内容
证明材料	30-50词	1-2个带数据的过往项目
时间线	15-25词	具体日期，拒绝模糊表述
报价	10-15词	若合适则给出清晰定价
行动号召（CTA）	15-20词	明确的下一步行动

总长度：195-285词（控制在300词以内最佳）

Anti-AI-Slop Rules

反AI套话规则

See

references/standout-strategies.md

for full guide. Critical rules:

DO	DON'T
Start with "Hey there" or use client's name	"Dear Hiring Manager" or "I hope this finds you well"
Mention 2 specifics from their job post	Generic "I'm the perfect fit for this role"
Use numbers: "23 similar projects", "$12K/month results"	Vague claims: "extensive experience"
Natural contractions: "I'll", "you're", "can't"	Robotic: "I would be delighted to"
One emoji max (👋 in greeting)	Multiple emojis or none at all
Specific timeline: "December 15th"	Vague: "as soon as possible"
Lead with solution approach	Lead with credentials

完整指南请见

references/standout-strategies.md

。核心规则：

要做	不要做
用“Hey there”或客户姓名开头	用“尊敬的招聘经理”或“祝您一切安好”开头
提及职位描述中的2个具体细节	泛泛而谈“我是这个职位的完美人选”
用数据说话：“23个类似项目”、“每月带来1.2万美元收益”	模糊表述：“丰富的经验”
使用自然缩写：“I'll”、“you're”、“can't”	机械表述：“我很荣幸能...”
最多使用1个表情符号（问候语中用👋）	多个表情符号或完全不用
具体时间线：“12月15日”	模糊表述：“尽快”
先讲解决方案思路	先讲资质

Testing Instruction Detection

测试指令识别

Clients often add tests like "Write HOWDY at the top" to filter spam. Always scan the job for:

Specific words to include
Questions to answer
Hidden instructions in the middle/end of posting

If found, follow them FIRST.

客户常加入类似“在顶部写HOWDY”的测试来过滤垃圾提案。务必全面扫描职位描述，查找：

需要包含的特定词汇
需要回答的问题
隐藏在职位描述中间/末尾的指令

若发现此类内容，优先执行。

Output Format

输出格式

Always provide TWO versions (user frequently asks for shorter):

markdown

undefined

始终提供两个版本（用户经常要求精简版）：

markdown

undefined

Proposal (Ready to Submit)

可直接提交的提案

[Full proposal following 7-part formula]

[遵循七段式公式的完整提案]

Shorter Version

精简版

[Condensed 100-150 word version]

[100-150词的浓缩版本]

One-Liner (For quick reference)

一句话电梯提案

[Single sentence pitch]

[单句推销话术]

Notes

备注

[Relevant past projects to mention]
[Questions to ask client]
[Red flags if any]

undefined

[可提及的相关过往项目]
[要向客户询问的问题]
[若有任何红色预警]

undefined

User's Style Preferences

用户风格偏好

From 12 Upwork conversations analyzed:

Greeting: "Hey there 👋" or "Hi there 👋"
Tone: Natural, humanized, conversational
Structure: Solution-first, then experience
Length: Always asks for shorter - start concise
Honesty: Never oversells, mentions learning areas
Tech mentions: n8n, Python, AI agents, Google Workspace

基于12次Upwork对话分析得出：

问候语：“Hey there 👋”或“Hi there 👋”
语气：自然、人性化、口语化
结构：先讲解决方案，再讲经验
长度：用户总是要求精简版——从简洁内容入手
诚实性：绝不夸大，提及学习领域
技术提及：n8n、Python、AI agents、Google Workspace

User's Skill Positioning

用户技能定位

See

references/user-profile.md

. Quick reference:

Highlight these:

n8n workflow automation
AI agents / Agentic AI (PIAIC certified)
Python, FastAPI
API integrations
Google Workspace automation
WhatsApp Business API

Be honest about:

React, Node.js (learning)
Make/Zapier (concepts understood, not hands-on)

详情请见

references/user-profile.md

。快速参考：

重点突出以下技能：

n8n工作流自动化
AI agents / Agentic AI（PIAIC认证）
Python、FastAPI
API集成
Google Workspace自动化
WhatsApp Business API

诚实告知以下情况：

React、Node.js（学习中）
Make/Zapier（理解概念，但无实操经验）

Engineering Methodology (Trust Signal)

工程方法论（信任信号）

User follows Spec-Driven Development (SDD) and Test-Driven Development (TDD).

When to mention this: Only for projects where quality/reliability matters:

Complex automation with multiple integrations
Long-term/ongoing projects
Clients who mention "reliable", "production-ready", "maintainable"
Higher-budget projects ($500+)
Clients burned by previous freelancers

DON'T mention for: Simple one-off tasks, quick fixes, low-budget gigs.

What you deliver (in repo):

```
specs/
```
- Requirements & design specs (per phase)
```
tests/
```
- Pytest test suites (written BEFORE code)
```
docs/ADRs/
```
- Architecture Decision Records
Self-documenting code with clear structure

How to phrase it naturally:

I follow spec-driven development — I'll document requirements first,
write tests, then implement. You'll get working code plus specs and
tests in the repo, not just "it works on my machine."

Shorter version:

I write specs and tests first, so you get reliable, documented code.

One-liner (when relevant):

All deliverables include specs, tests, and documentation in the repo.

See

references/engineering-methodology.md

for full details.

用户遵循Spec-Driven Development (SDD，规范驱动开发)和Test-Driven Development (TDD，测试驱动开发)。

**何时提及：**仅适用于对质量/可靠性有要求的项目：

涉及多集成的复杂自动化项目
长期/持续进行的项目
客户提及“可靠”、“生产就绪”、“可维护”的项目
高预算项目（500美元以上）
曾被前自由职业者坑过的客户

**不要提及的场景：**简单一次性任务、快速修复、低预算项目。

交付内容（在代码库中）：

```
specs/
```
- 各阶段的需求与设计规范
```
tests/
```
- Pytest测试套件（在编写代码前完成）
```
docs/ADRs/
```
- 架构决策记录
自文档化代码，结构清晰

自然表述方式：

我遵循规范驱动开发——先记录需求，再编写测试，最后实现功能。您将获得可运行的代码，以及代码库中的规范和测试文档，而不是仅能在我本地运行的代码。

精简版：

我先写规范和测试，确保您获得可靠、有文档记录的代码。

相关场景下的一句话表述：

所有交付内容均包含代码库中的规范、测试和文档。

详情请见

references/engineering-methodology.md

。

Example Transformation

示例对比

Generic AI slop:

Dear Hiring Manager,
I hope this message finds you well. I am writing to express my keen interest in your project. With my extensive experience in automation and workflow development, I am confident I can deliver exceptional results. I have worked on numerous similar projects and would be delighted to discuss further.

Winning proposal:

Hey there 👋

I can help you automate your Google Sheets → AI → PDF → Email workflow in n8n.

Here's the plan:
- Connect Google Sheets with scheduled triggers (twice weekly)
- Pass data to ChatGPT/DeepSeek for processing
- Generate formatted PDF (via API or Google Docs conversion)
- Auto-send via Gmail

I recently built a similar workflow connecting Sheets with ChatGPT for automated reporting — happy to share details.

When would you like to kick off?

通用AI生成的垃圾提案：

Dear Hiring Manager,
I hope this message finds you well. I am writing to express my keen interest in your project. With my extensive experience in automation and workflow development, I am confident I can deliver exceptional results. I have worked on numerous similar projects and would be delighted to discuss further.

优质提案：

Hey there 👋

I can help you automate your Google Sheets → AI → PDF → Email workflow in n8n.

Here's the plan:
- Connect Google Sheets with scheduled triggers (twice weekly)
- Pass data to ChatGPT/DeepSeek for processing
- Generate formatted PDF (via API or Google Docs conversion)
- Auto-send via Gmail

I recently built a similar workflow connecting Sheets with ChatGPT for automated reporting — happy to share details.

When would you like to kick off?

upwork-proposal

Original

Translation

Upwork Proposal Skill

Upwork提案撰写技能

The Problem We're Solving

我们解决的问题

Key Stats (2025-2026)

关键数据（2025-2026）

Workflow

工作流程

Required Input

必填输入

Optional Inputs (Ask if relevant)

可选输入（按需询问）

The 7-Part Formula (24%+ Response Rate)

七段式提案公式（回复率提升24%+）

Anti-AI-Slop Rules

反AI套话规则

Testing Instruction Detection

测试指令识别

Output Format

输出格式

Proposal (Ready to Submit)

可直接提交的提案

Shorter Version

精简版

One-Liner (For quick reference)

一句话电梯提案

Notes

备注

User's Style Preferences

用户风格偏好

User's Skill Positioning

用户技能定位

Engineering Methodology (Trust Signal)

工程方法论（信任信号）

Example Transformation

示例对比

Checklist Before Output

输出前检查清单