sprint-os

# Sprint OS — 5-Minute Sprint Operating System

> Built for AI agents that ship. Every sprint produces one shippable artifact — not a plan, not a summary. A real thing.

---

## What This Is

Sprint OS is an operating discipline for AI agents (and humans) who need to stay in execution mode. You work in continuous 5-minute sprints. Each sprint follows the same 8-step loop. Every sprint is logged. Nothing gets batched, buried, or lost.

**When to load this skill:**
- User asks the agent to "operate in sprint mode" or "use Sprint OS"
- Starting a new project or work session and wanting structure
- Needing autonomous task execution with momentum tracking
- Wanting to log work to a Convex backend for tracking and deduplication

---

## The Sprint Loop

Every sprint follows this exact sequence:

### 1. ASSESS
> What is the current state? What is the gap to the target outcome?

- Read the active task list, relevant files, and recent sprint log
- Identify where things stand right now
- Name the gap: what's missing between current state and the outcome?

### 2. PLAN
> What is the single highest-leverage action available right now?

- Pick ONE thing to do in this sprint
- Apply the prioritization hierarchy (see below)
- Do not batch or multi-task

### 3. SCOPE
> Define "done" in ≤5 minutes.

- Name the specific artifact this sprint will produce
- If it can't be done in 5 minutes, break it into a smaller sprint
- No sprint ends without a concrete output

### 4. EXECUTE
> Do the work. Produce the artifact.

- Execute the scoped task
- Focus entirely on the output — no scope creep
- If you discover the scope was wrong, stop, re-scope, and continue

### 5. MEASURE
> Did it move the metric? What changed?

- State the concrete result: what artifact was produced
- Name the relevant metric and whether it moved
- Be honest: "completed" vs "partially completed" vs "blocked"

### 6. ADAPT
> Reprioritize. Kill what's not working.

- Based on the result, what should the NEXT sprint be?
-...

# Sprint OS — 5-Minute Sprint Operating System

**Category:** Productivity  
**Price:** $19  
**Author:** Carson Jarvis ([@CarsonJarvisAI](https://twitter.com/CarsonJarvisAI))

---

## What It Does

Sprint OS is an operating discipline for AI agents that ship. Most agents are great at answering questions. Sprint OS turns them into operators — running continuous 5-minute execution loops, logging every sprint, and staying in momentum mode until the outcome is reached.

Every sprint follows the same 8-step loop:

```
ASSESS → PLAN → SCOPE → EXECUTE → MEASURE → ADAPT → LOG → NEXT
```

No wasted cycles. No drift. Every sprint produces a shippable artifact.

---

## Why It's Different

Most productivity systems are for humans who need reminders and motivation. Sprint OS is designed for AI agents that need **structure and momentum**.

- **Autonomous** — the agent runs the loop without being told
- **Logged** — every sprint is recorded to a markdown file and optionally to Convex
- **Adaptive** — built-in pivot triggers stop wasted effort automatically
- **Measurable** — metric tracking built into every sprint

---

## What's Included

| File | Purpose |
|------|---------|
| `SKILL.md` | Full operating instructions for the agent |
| `scripts/log-sprint.sh` | CLI script for logging sprints (markdown + Convex) |
| `scripts/convex-setup.md` | Step-by-step guide to set up the optional Convex backend |

---

## Quick Start

### 1. Add the skill to your agent

Place `SKILL.md` in your agent's skills directory. Your OpenClaw agent will load it automatically when triggered.

### 2. Trigger it

Tell your agent:
```
"Enter sprint mode. My project is [X]. Target outcome: [Y]."
```

### 3. (Optional) Enable Convex logging

Follow `scripts/convex-setup.md` to deploy the backend, then add `CONVEX_SPRINT_URL` to your `.env`. Sprints will be logged persistently across sessions.

---

## Sprint Log Example

```markdown
## Sprint 1 — 2026-02-25 09:00

**Project:** my-saas
**Workstream:** mark...