Files
antigravity-skills-reference/skills/writing-skills/references/tier-3-platform

description, metadata
description metadata
When to use Tier 3 (Platform) skill architecture for large platforms.
tags
tier-3
platform
enterprise
cloudflare-pattern

Tier 3: Platform Skills

Enterprise-grade skills for entire platforms (AWS, Cloudflare, Convex, etc).

When to Use

  • Entire platform: 10+ products/services
  • 1000+ lines total: Would overwhelm context if monolithic
  • Complex decision logic: Users start with "I need X" not "I want product Y"
  • Undocumented gotchas: Tribal knowledge is critical

The Cloudflare Pattern

Based on cloudflare-skill by Dillon Mulroy.

Structure

my-platform/
├── SKILL.md                  # Decision trees only
└── references/
    └── <product>/
        ├── README.md         # Overview, when to use
        ├── api.md            # Runtime API reference
        ├── configuration.md  # Config options
        ├── patterns.md       # Usage patterns
        └── gotchas.md        # Pitfalls, limits

The 5-File Pattern

Each product directory has exactly 5 files:

File Purpose When to Load
README.md Overview, when to use Always first
api.md Runtime APIs, methods Implementing features
configuration.md Config, environment Setting up
patterns.md Common workflows Best practices
gotchas.md Pitfalls, limits Debugging

Decision Trees

The power of Tier 3 is decision trees that help the AI choose:

Need to store data?
├─ Simple key-value → kv/
├─ Relational queries → d1/
├─ Large files/blobs → r2/
├─ Per-user state → durable-objects/
└─ Vector embeddings → vectorize/

Slash Command Integration

Create a slash command to orchestrate:

---
description: Load platform skill and get contextual guidance
---

## Workflow

1. Load skill: `skill({ name: 'my-platform' })`
2. Identify product from decision tree
3. Load relevant reference files based on task

| Task | Files |
|------|-------|
| New setup | README.md + configuration.md |
| Implement feature | api.md + patterns.md |
| Debug issue | gotchas.md |

Progressive Disclosure in Action

  • Startup: Only name + description (~100 tokens)
  • Activation: SKILL.md with trees (<5000 tokens)
  • Navigation: One product's 5 files (as needed)

Result: 60+ product references without blowing context.

Checklist

  • SKILL.md contains ONLY decision trees + index
  • Each product has exactly 5 files
  • Decision trees cover all "I need X" scenarios
  • Cross-references stay one level deep
  • Slash command created for orchestration
  • Every product has gotchas.md