hackathon-submission-prep

Goal

Compile and validate all required hackathon submission artifacts into a complete, polished submission package ready for upload.

---

Trigger Conditions

Use this skill when:

• The demo video is uploaded and the pitch deck is finalized
• Submission deadline is approaching and artifacts must be validated
• A submission platform (Devpost, Dorahacks, etc.) requires a structured description
• The team needs a checklist of what is complete versus what is still missing
• Invoked during Phase 8; submit output to the platform at least 1 hour before the deadline

---

Inputs

Input	Type	Required	Description
project_title	string	Yes	Name of the project
tagline	string	Yes	One-sentence description
problem_statement	string	Yes	The problem being solved
solution_summary	string	Yes	How the project solves it
tech_stack	string[]	Yes	Technologies used
team_members	object[]	Yes	Team member names and roles
demo_url	string	No	Live demo URL
video_url	string	No	Demo video URL
repo_url	string	Yes	Source code repository URL
track	string	Yes	Hackathon track being entered
submission_platform	string	No	Platform name (e.g., Devpost, Dorahacks)
custom_fields	object[]	No	Platform-specific required fields

---

Outputs

Output	Description
submission_description	Full formatted project description for submission
short_description	280-character tagline/elevator pitch
artifact_checklist	All required submission items with completion status
missing_artifacts	Items not yet provided that must be completed
submission_quality_score	1–10 rating of submission readiness
last_mile_actions	Final actions required before submitting

---

Rules

1. submission_description must include: problem, solution, tech stack, team, and demo link.
2. short_description must be 280 characters or fewer.
3. artifact_checklist must verify: repo URL accessible, README exists, demo video linked, team listed.
4. Assign submission_quality_score using: completeness (40%), clarity (30%), impact framing (30%).
5. last_mile_actions must be ordered by priority with estimated time per action.
6. Flag any missing required field from custom_fields as a blocking missing_artifact.
7. Do not submit placeholder text; flag any [TBD] fields as incomplete in missing_artifacts.

---

Output Format

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>

---

Example

Input:

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"

Output:

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

---

Context Files

Knowledge Base

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

Playbooks

• playbooks/hackathon-workflow.md