An enterprise prototyping operating model — a GitHub-based template repo that any organization can fork to channel prototyping energy into structured value creation. NOT an app. A system.
The core thesis: people in enterprises are prototyping constantly (with AI tools, on weekends, in shadows). You can't stop it. You can channel it. This repo IS the channel.
This is a template repository, not a product. The value is in:
- Documentation — the operating model, team manifests, guidelines
- GitHub native workflows — issue templates, PR templates, labels, project boards
- Submission templates — standardized format for prototype submissions (as PRs)
- Evaluation rubrics — how to score and select prototypes
- Claude Code extensions — agents/skills for submission review, evaluation, grouping, dashboard generation
- Simple frontend — a single-page dashboard to visualize submissions, status, and outcomes
- No Jira, no ServiceNow, no custom app
- Git is the universal tool engineers already use
- PRs = submissions (decisions recorded, challenged, approved/rejected)
- Issues = tracking evaluation and graduation
- Labels = categorization and status
- GitHub Projects = pipeline visualization
- Everything is version controlled, auditable, in one place
- Submissions must include a WORKING prototype (minimum viable demo)
- Not sketches or slide decks (too early)
- Not deployed applications (too late)
- The sweet spot: "I built this thing, it works, here's what it solves"
- People submit at their own pace (not one-day hackathon pressure)
- Quarterly evaluation cycles by a cross-functional review panel
- Top 5 rewarded, top 15 invested in
prototype-ops/
├── CLAUDE.md
├── REQUIREMENTS.md
├── README.md # What this repo is, how to fork and use it
├── docs/
│ ├── operating-model.md # The full operating model (5 phases)
│ ├── team-manifest/
│ │ ├── leadership-team.md # Do's and don'ts for leadership
│ │ ├── organising-team.md # Do's and don'ts for organising team
│ │ └── submitters-guide.md # Guide for people submitting prototypes
│ ├── guidelines/
│ │ ├── approved-tools.md # AI tools offered by the company
│ │ ├── security-guidelines.md # What you can and can't use/share
│ │ ├── problem-framing.md # How to think about problems (dollar value)
│ │ ├── effort-estimation.md # How to estimate build effort
│ │ └── value-creation.md # How to articulate value created
│ ├── evaluation/
│ │ ├── rubric.md # Scoring criteria for submissions
│ │ ├── graduation-paths.md # 4 paths: graduate, iterate, shelve, kill
│ │ └── review-process.md # How the quarterly review works
│ └── references/
│ ├── org-level-problems.md # Example org-level problems to solve
│ └── past-submissions.md # Template for documenting past cycles
├── templates/
│ ├── submission-template.md # What a PR-based submission looks like
│ ├── evaluation-template.md # Reviewer scoring template
│ └── graduation-template.md # Post-evaluation outcome template
├── .github/
│ ├── ISSUE_TEMPLATE/
│ │ ├── submission.yml # GitHub issue form for submissions
│ │ ├── evaluation.yml # Issue form for evaluation tracking
│ │ └── graduation.yml # Issue form for graduation decisions
│ ├── PULL_REQUEST_TEMPLATE.md # PR template for prototype submissions
│ ├── labels.yml # Predefined labels for categorization
│ └── workflows/
│ ├── submission-check.yml # Validate submission format
│ └── quarterly-report.yml # Generate quarterly summary
├── .claude/
│ ├── settings.json
│ ├── agents/
│ │ └── reviewer.md # Agent that reviews submissions
│ └── skills/
│ ├── evaluate-submission.md # Skill to score a submission against rubric
│ ├── group-submissions.md # Skill to identify overlapping/similar submissions
│ ├── quarterly-report.md # Skill to generate cycle summary report
│ └── dashboard-update.md # Skill to update the frontend dashboard
├── dashboard/ # Simple static frontend
│ ├── index.html # Single-page dashboard
│ ├── styles.css # Monochrome neo-brutalist styling
│ └── data/
│ └── submissions.json # Data file generated from GitHub issues/PRs
└── examples/
├── example-submission-1.md # Example: AI-powered inventory alerting
├── example-submission-2.md # Example: Meeting summary automation
└── example-submission-3.md # Example: Customer feedback classifier
- Documentation: Markdown (everything is markdown)
- Workflows: GitHub Actions
- Dashboard: Static HTML/CSS/JS (no framework, no build step — forkable simplicity)
- Extensions: Claude Code agents and skills
- Styling: Monochrome neo-brutalist (from global design system)
- Direct, practical, no corporate fluff
- Written for the person who has to actually set this up
- Opinionated — this repo takes positions, doesn't try to please everyone
- Examples are concrete, not abstract
- Small, focused commits
- Use conventional commit prefixes: feat:, docs:, chore:, fix:
- Never commit to main directly — use feature branches
- Reference issue numbers in commits and PRs
- Not a product with user auth, databases, multi-tenancy
- Not a hackathon management tool
- Not an AI governance framework (though it helps with shadow AI)
- Not theoretical — everything should be immediately usable when forked