Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Draft: Module 3 - Intent-Driven Development

Writing architectural intent, adversarial critique, and evidence-based validation.

Overview

This module teaches intent-driven development with Sruja—expressing what you want the architecture to achieve, then validating that reality matches intent.

Lessons

Lesson 1: Writing Architectural Intent

Expressing desired outcomes as formal intent

  • Intent syntax and structure
  • Outcome-driven requirements
  • Linking intent to evidence
  • Intent provenance tracking

Lesson 2: Critique Engine & Adversarial Review

Let AI challenge your architecture decisions

  • The "Angry Agent" critique engine
  • Running adversarial review
  • Interpreting critique results
  • Responding to critique

Lesson 3: Evidence Mapping & Compliance

Prove your architecture meets intent

  • Evidence collection
  • Intent validation against code
  • Compliance reporting
  • Drift detection and resolution

Learning Outcomes

  • ✅ Write formal architectural intent
  • ✅ Use critique engine to challenge decisions
  • ✅ Map evidence from code to intent
  • ✅ Validate compliance in CI/CD

Prerequisites

  • Completed Advanced Architects Modules 1-2
  • Familiarity with Sruja policies and contracts
  • Understanding of CI/CD workflows

Estimated Time

2-3 hours

Raw Thoughts from Analysis

v0.34.0, v0.35.0 Features to Cover:

  • intent command and domain
  • Phase 4: Adversarial Critique Engine ("Angry Agent")
  • Phase 3: Dynamic intent check and evidence mapping
  • sruja propose command
  • sruja review command
  • Intent-first development
  • Architectural rationale (ADR)
  • Critique-driven iteration
  • Evidence-based validation
  • Compliance and audit trails

Key Commands:

  • sruja intent propose -r . --from-diff
  • sruja intent check -r .
  • sruja critique -r .
  • sruja review -r .
  • sruja propose -r .

Lesson Ideas:

  1. Writing Architectural Intent

    • Intent syntax: what you want, not how
    • Outcome vs constraint intent
    • Linking to ADR/rationale
    • Example: "API latency must be < 200ms" vs "We use caching"

  2. Critique Engine

    • The adversarial agent ("Angry Agent")
    • What it challenges: assumptions, trade-offs, risks
    • Running sruja critique
    • Interpreting critique report

  3. Evidence Mapping

    • Collecting evidence from code
    • Mapping to intent requirements
    • Validation reporting
    • Drift detection