Skip to main content

Core Concepts

Six building blocks of jaan.to.


Skills

Commands that generate outputs.

  • Pattern: /role-domain-action
  • Example: /jaan-to:pm-prd-write
  • Output: Markdown files in jaan-to/outputs/

Skills follow a two-phase workflow:

  1. Analysis - Read context, ask questions
  2. Generation - Create artifact, preview, write

Learn more


Stacks

Your team and tech context.

FileContains
jaan-to/context/tech.mdLanguages, frameworks, tools
jaan-to/context/team.mdTeam size, ceremonies, norms
jaan-to/context/integrations.mdJira, GitLab, Slack setup

Skills read context to generate context-aware outputs.

Learn more


Templates

Output formats for outputs.

  • Location: jaan-to/templates/{skill}.template.md
  • Contains: Section structure, placeholders
  • Override: Create your own version

Skills fill templates with generated content.


Learning

System improves from feedback.

Each skill has a LEARN.md file with four categories:

CategoryExample
Better Questions"Ask about i18n requirements"
Edge Cases"Check for multi-tenant scenarios"
Workflow"Generate metrics JSON alongside PRD"
Common Mistakes"Don't assume single region"

Add lessons: /jaan-to:learn-add {skill} "lesson text"

Learn more


Hooks

Automated triggers before/after actions.

TypeWhenExample
PreToolUseBefore writeValidate required sections
PostToolUseAfter writePrompt for feedback

Hooks run automatically. No action needed.

Learn more


Guardrails

Safety boundaries.

  • Safe paths: Only write to jaan-to/
  • Preview first: Always show before saving
  • Approval required: You confirm every write

Guardrails cannot be disabled.

Learn more