hackathon-submission-prep

hackathon-submission-prep

目标

汇总并验证所有必需的黑客松提交材料，整理成完整、精致、可供上传的提交包。

---

触发条件

在以下情况下使用此技能：

• 演示视频已上传，路演幻灯片已定稿
• 提交截止时间临近，需要验证材料
• 提交平台（Devpost、Dorahacks 等）要求填写结构化说明
• 团队需要核对清单，了解哪些已完成、哪些仍待补齐
• 在第 8 阶段调用；至少提前 1 小时将输出提交到平台

---

输入

输入	类型	必填	描述
project_title	字符串	是	项目名称
tagline	字符串	是	一句话描述
problem_statement	字符串	是	要解决的问题
solution_summary	字符串	是	项目如何解决该问题
tech_stack	字符串[]	是	使用的技术
team_members	对象[]	是	团队成员姓名和角色
demo_url	字符串	否	在线演示 URL
video_url	字符串	否	演示视频 URL
repo_url	字符串	是	源代码仓库 URL
track	字符串	是	参与的黑客松赛道
submission_platform	字符串	否	平台名称（例如 Devpost、Dorahacks）
custom_fields	对象[]	否	平台特定的必填字段

---

输出

输出	描述
submission_description	用于提交的完整格式化项目说明
short_description	280 字符以内的标语／简短介绍
artifact_checklist	所有必需提交项及其完成状态
missing_artifacts	尚未提供、必须补齐的项目
submission_quality_score	对提交准备情况的 1–10 分评级
last_mile_actions	提交前所需的最后操作

---

规则

1. submission_description 必须包括：问题、解决方案、技术栈、团队和演示链接。
2. short_description 不得超过 280 个字符。
3. artifact_checklist 必须核实：仓库 URL 可访问、README 存在、已添加演示视频链接、已列出团队成员。
4. 按以下权重计算 submission_quality_score：完整性（40%）、清晰度（30%）、影响力阐述（30%）。
5. last_mile_actions 必须按优先级排序，并为每项操作估算所需时间。
6. 将 custom_fields 中缺少的任何必填字段标记为阻塞性 missing_artifact。
7. 不要提交占位文本；将任何 [TBD] 字段标记为 missing_artifacts 中未完成的项目。

---

输出格式

submission_description: |
  <Full Markdown submission text>

short_description: "<≤280 characters>"

artifact_checklist:
  - item: "<artifact name>"
    status: "<complete|incomplete|missing>"
    url_or_value: "<link or value if complete>"

missing_artifacts:
  - item: "<name>"
    blocking: <true|false>
    action_required: "<what to do>"

submission_quality_score:
  score: <1-10>
  breakdown:
    completeness: <1-10>
    clarity: <1-10>
    impact_framing: <1-10>

last_mile_actions:
  - priority: <number>
    action: "<what to do>"
    estimated_minutes: <number>

---

示例

输入：

project_title: "AnchorAI"
tagline: "An AI companion that remembers your emotional journey so you never have to start over."
problem_statement: "College students wait 6 weeks for counseling while anxiety compounds daily."
solution_summary: "GPT-4 with Redis-backed session memory for always-available, context-aware emotional support."
tech_stack: ["Python", "FastAPI", "React", "Redis", "OpenAI GPT-4"]
team_members:
  - name: "Alex"
    role: "Backend"
  - name: "Sam"
    role: "Frontend"
  - name: "Jordan"
    role: "Design"
repo_url: "https://github.com/team/anchorai"
track: "AI for Mental Health"
submission_platform: "Devpost"

输出：

submission_description: |
  ## Inspiration
  40% of college students report anxiety as their top health concern, yet the average campus counseling wait is 6 weeks. Most students give up before they're seen.

  ## What It Does
  AnchorAI is a GPT-4 powered emotional support companion that remembers your history across sessions. It opens every conversation with context from your last check-in — so you never have to re-explain your situation.

  ## How We Built It
  Python FastAPI backend with OpenAI GPT-4 API. Session summaries stored in Redis and injected into the system prompt. React frontend with a clean single-page chat UI.

  ## Challenges
  Getting memory injection to stay within GPT-4 context window limits required hard-capping summaries at 150 tokens and testing multiple summarization prompts.

  ## Accomplishments
  Working live demo with genuine session memory recall. Crisis escalation card renders on distress detection.

  ## What We Learned
  The emotional framing of AI responses matters more than technical sophistication. Users notice tone before they notice features.

  ## What's Next
  Human counselor escalation. University licensing. Multi-session longitudinal mood tracking.

short_description: "AnchorAI: GPT-4 emotional support companion that remembers your mental health journey. No waitlist. No starting over. Always there. 🧠"

artifact_checklist:
  - item: "Source code repository"
    status: "complete"
    url_or_value: "https://github.com/team/anchorai"
  - item: "Demo video"
    status: "complete"
    url_or_value: "https://youtube.com/watch?v=demo"
  - item: "Live demo URL"
    status: "incomplete"
    url_or_value: ""
  - item: "README with setup instructions"
    status: "complete"
    url_or_value: "https://github.com/team/anchorai#readme"

missing_artifacts:
  - item: "Live demo URL"
    blocking: false
    action_required: "Deploy to Railway or Render; update Devpost submission with URL"

submission_quality_score:
  score: 8
  breakdown:
    completeness: 9
    clarity: 8
    impact_framing: 8

last_mile_actions:
  - priority: 1
    action: "Deploy live demo to Railway and update Devpost URL field"
    estimated_minutes: 20
  - priority: 2
    action: "Add 3 screenshots to Devpost gallery"
    estimated_minutes: 10
  - priority: 3
    action: "Confirm repo is public and README loads correctly"
    estimated_minutes: 5

---

上下文文件

知识库

• ../hackathon-shared-resources/knowledge/hackathon-submission-guidelines.md
• ../hackathon-shared-resources/knowledge/hackathon-winning-patterns.md
• ../hackathon-shared-resources/knowledge/hackathon-pitch-strategy.md

操作手册

• ../hackathon-shared-resources/playbooks/hackathon-workflow.md