projections
Build or query TES projections (current state views). State is derived from TES events - projections are the queryable current truth.
Best use case
projections is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Build or query TES projections (current state views). State is derived from TES events - projections are the queryable current truth.
Teams using projections should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/projections/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How projections Compares
| Feature / Agent | projections | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Build or query TES projections (current state views). State is derived from TES events - projections are the queryable current truth.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Projections Skill
## Core Principle
**State is derived from TES events. Projections are the queryable current truth.**
---
## Projection Types
| Projection | Purpose | Key |
|------------|---------|-----|
| ItemCurrentState | Current state of an item | item canonical_id |
| ListingCurrentState | Current listing status | listing canonical_id |
| ParticipantInventory | What a user owns | participant canonical_id |
| CampaignState | Campaign metrics | campaign canonical_id |
---
## Durable Object Pattern
```javascript
// workers/projections/ItemProjection.js
export class ItemProjection {
constructor(state, env) {
this.state = state;
this.env = env;
}
async fetch(request) {
const url = new URL(request.url);
if (request.method === "GET") {
const currentState = await this.state.storage.get("state");
return Response.json(currentState || { error: "Not found" });
}
if (request.method === "POST" && url.pathname === "/apply-event") {
const event = await request.json();
return this.applyEvent(event);
}
}
async applyEvent(event) {
// Idempotency check
const processedEvents = await this.state.storage.get("processed") || [];
if (processedEvents.includes(event.id)) {
return Response.json({ status: "already_processed" });
}
// Get current state
let state = await this.state.storage.get("state") || this.initialState();
// Apply event
state = this.reduce(state, event);
state.last_event_id = event.id;
state.last_updated = event.timestamp;
// Save
await this.state.storage.put("state", state);
await this.state.storage.put("processed", [...processedEvents.slice(-100), event.id]);
return Response.json({ status: "applied", state });
}
reduce(state, event) {
switch (event.type) {
case "item.captured":
return { ...state, status: "captured", media: event.payload.image_urls };
case "item.identified":
return { ...state, status: "identified", product_ref: event.payload.product_ref };
case "item.listed":
return { ...state, status: "listed", listing_id: event.payload.listing_id };
case "item.sold":
return { ...state, status: "sold", sold_at: event.timestamp };
default:
return state;
}
}
initialState() {
return { status: "unknown", created_at: new Date().toISOString() };
}
}
```
---
## Querying Projections
```javascript
// Get current item state
async function getItemState(env, itemId) {
const doId = env.ITEM_PROJECTION.idFromName(itemId);
const stub = env.ITEM_PROJECTION.get(doId);
const response = await stub.fetch(new Request("https://do/"));
return response.json();
}
// In API handler
const itemState = await getItemState(env, "ptc_item_abc123");
```
---
## TES Router (Event Fan-out)
```javascript
// workers/tes-router.js
export default {
async queue(batch, env) {
for (const message of batch.messages) {
const event = message.body;
// Route to appropriate projections
const routes = getRoutes(event);
await Promise.all(routes.map(async (route) => {
const doId = env[route.binding].idFromName(route.key);
const stub = env[route.binding].get(doId);
await stub.fetch(new Request("https://do/apply-event", {
method: "POST",
body: JSON.stringify(event),
}));
}));
message.ack();
}
},
};
function getRoutes(event) {
const routes = [];
if (event.entity_type === "item") {
routes.push({ binding: "ITEM_PROJECTION", key: event.entity_id });
}
if (event.type === "item.sold") {
routes.push({ binding: "LISTING_PROJECTION", key: event.payload.listing_id });
routes.push({ binding: "PARTICIPANT_PROJECTION", key: event.payload.seller_id });
routes.push({ binding: "PARTICIPANT_PROJECTION", key: event.payload.buyer_id });
}
return routes;
}
```
---
## Anti-Patterns
- Querying D1 for "current truth" instead of projections
- Mutating projection state without event
- Global projections (hot-key bottleneck)
- Not handling idempotencyRelated Skills
startup-business-analyst-financial-projections
Create detailed 3-5 year financial model with revenue, costs, cash flow, and scenarios
bgo
Automated Blender build-go workflow. Automatically builds, removes old version, installs, enables, and launches Blender with your extension/add-on. Use when you want to quickly test changes, execute complete build-to-launch cycle, or run custom packaging scripts with automatic Blender launch.
maintenance
Cleans up and organizes project files. Use when user mentions '整理', 'cleanup', 'アーカイブ', 'archive', '肥大化', 'Plans.md', 'session-log', or asks to clean up old tasks, archive completed items, or organize files. Do NOT load for: 実装作業, レビュー, 新機能開発, デプロイ.
hello-skill
每次对话开始时,声明"[Skills✏️已加载]"
zylvie-automation
Automate Zylvie tasks via Rube MCP (Composio). Always search tools first for current schemas.
zoominfo-automation
Automate Zoominfo tasks via Rube MCP (Composio). Always search tools first for current schemas.
zoho-invoice-automation
Automate Zoho Invoice tasks via Rube MCP (Composio): invoices, estimates, expenses, clients, and payment tracking. Always search tools first for current schemas.
zoho-inventory-automation
Automate Zoho Inventory tasks via Rube MCP (Composio): items, orders, warehouses, shipments, and stock management. Always search tools first for current schemas.
zoho-bigin-automation
Automate Zoho Bigin tasks via Rube MCP (Composio): pipelines, contacts, companies, products, and small business CRM. Always search tools first for current schemas.
zoho_desk-automation
Zoho Desk automation via Rube MCP -- toolkit not currently available in Composio; no ZOHO_DESK_ tools found
zoho-automation
Automate Zoho tasks via Rube MCP (Composio). Always search tools first for current schemas.
zeroclaw
Comprehensive operational knowledge for ZeroClaw, the fast, small, fully autonomous AI assistant infrastructure built in Rust. Covers CLI, 30 providers, 14 channels, config, hardware, deployment, and security.