The Intent Playbook
Not courses. Not theory. Executable blueprints.
Go from zero to your first Agent-Ready Spec in 10 minutes.
A Product Intent Loop for Anthropic's AI-Native SDLC Playbook
Anthropic's AI-native SDLC playbook accelerates execution. This companion explains the product intent loop that feeds it: Sense, Decide, Specify, and Learn.
What Is Intent Engineering?
The 3-minute primer on intents, intent specs, and why 'intent engineering' is a discipline — not a buzzword.
The 8-Part IntentSpec
The periodic table of Intent Engineering. Every executable spec is made of these 8 parts. Memorize them.
The Anatomy of an Agent-Ready Spec
Learn what makes a specification truly 'executable' by AI agents like Cursor and Windsurf. This guide breaks down the key components.
Why Specs Fail: The 5 Anti-Patterns
Most specifications are destined for the trash can because they are written for humans to ignore, not for agents to execute. Here are the 5 traps to avoid.
Your First Spec in 10 Minutes
You don't need fancy tools to start Intent Engineering. Open a text editor and write your first agent-ready spec — objective, outcomes, constraints, and the verification that proves it worked.
Your First Intent in Pathmode
A step-by-step walkthrough of writing your first intent spec in Pathmode — from capturing friction to handing the spec to your AI agent.
Evidence 101: The Discovery Play
Evidence is not a deliverable, it is a record of judgment. Learn how to capture user signals, review them one at a time, and use them as backing for exact claims in an Intent Spec.
From User Interview to Reviewed Evidence
Your interview notes are full of signal. Learn how to turn raw conversations into structured signals you can review, and use as backing for the exact claims in a spec.
Prioritizing Your Intent Backlog
Stop prioritizing by gut feel. Use evidence weight to decide which intent to spec next — and which to kill.
The Intent Spec as Alignment Artifact
Stop writing PRDs to align your team. An intent spec is shorter, testable, and eliminates the 'that's not what I meant' conversation.
From Support Ticket to Shipped Feature
A step-by-step worked example showing how a raw support ticket becomes a shipped feature using the Pathmode workflow.
Spec Template: Login Flow
A ready-to-use Intent Specification template for implementing a standard user authentication flow. Copy and adapt for your project.
Template: SaaS Onboarding Friction Patterns
A catalogue of the most common friction patterns in B2B SaaS onboarding, organized by stage. Use it as a checklist for what to look for in your own data.
The Product Constitution: Rules That Outlive a Single Spec
The same constraint, retyped into thirty specs, eventually gets retyped wrong. The Constitution is where a rule lives once — with its reason, its blast radius, and its teeth — and rides along on every spec the agent receives.
Verification Is a Feedback Loop, Not a Test List
The agent said "done" and the tests were green. A week later, checkout was quietly down 4%. A test list tells you the code runs. A feedback loop tells you the intent held — here are the four legs that catch what the tests can't.
When the Build Contradicts the Spec
The agent built the feature and learned the spec was wrong. That correction usually dies in a PR comment, so the next agent rebuilds the same dead premise. A finding is how the build writes back, a proposal is how it names the exact fix, and a person accepting it is how the loop finally converges.
Wire Pathmode Into Your Coding Agent
Your agent already writes good code. It builds the wrong thing because the judgment lives somewhere it cannot see. Wire it to the Pathmode MCP server so Claude Code, Cursor, and Codex preflight the repo's intent.md, read the constitution and the evidence, wait for a person to authorize the revision, and hand the pull request back to be graded.
A Product Intent Loop for Anthropic's AI-Native SDLC Playbook
Anthropic's AI-native SDLC playbook accelerates execution. This companion explains the product intent loop that feeds it: Sense, Decide, Specify, and Learn.
Want the latest thinking on intent engineering? Read The Builder's Log — our blog on product discovery and AI-native teams.