Explain

Share scenarios they've probably experienced:

• "Claude suggested a perfectly reasonable solution... that conflicted with a design decision you made 3 months ago"
• "Claude built a feature that works in isolation but doesn't fit your existing patterns"
• "Every new session, you're re-explaining your architecture"
• "Claude invented an abstraction you now have to maintain, when a simpler approach would have worked"

The root cause: Claude has vast general knowledge but zero knowledge of YOUR specific system. It fills the gaps with reasonable-sounding guesses.

2. The Solution (Keep It Simple)

Structured context = telling Claude about your system BEFORE it starts making decisions.

It's like onboarding a new developer:

• Here's what we're building
• Here's how it's structured
• Here are the patterns we use (and avoid)
• Here are the constraints you need to know

The key insight: This isn't about making Claude smarter. It's about giving it the information it needs to apply its intelligence appropriately.

3. What Goes in Structured Context

The essentials:

• System Overview: What does this do? Who uses it?
• Architecture: How is it structured? What are the key components?
• Tech Stack: What specific technologies and why?
• Patterns: What patterns do we use? What do we avoid?
• Constraints: What limits must Claude respect?

4. Before/After Example

Without structured context:

User: "Add a caching layer for the user lookup"

Claude: "I'll add Redis caching using the popular redis-om library
with a 24-hour TTL..."

[Problem: You use Memcached, you have a 1-hour max TTL policy,
and you have an existing caching utility Claude didn't know about]

With structured context:

CLAUDE.md:
- Cache: Memcached via /lib/cache utility
- Constraint: Max TTL 1 hour (compliance requirement)
- Pattern: All caching goes through CacheService

User: "Add a caching layer for the user lookup"

Claude: "I'll add caching using your existing CacheService in /lib/cache.
Based on your 1-hour max TTL constraint, I'll set TTL to 3600 seconds.
Here's the implementation following your existing patterns..."

[Claude works WITHIN your system instead of inventing a new one]

5. How to Get Started

1. Run /scs-vibe:init - 15 minute setup
2. Review and tweak - Add anything specific
3. Commit it - Version control your context
4. Run /scs-vibe:validate occasionally - Keep it fresh

6. Address Common Objections

"I already have a CLAUDE.md" Great! Run /scs-vibe:validate to check if it's still accurate. Context gets stale fast.

"This seems like a lot of work" 15 minutes. That's it. /scs-vibe:init does most of the work - you're just confirming and adding detail.

"Will Claude actually use this?" Yes. Claude reads CLAUDE.md at the start of every session. The more specific you are, the better it works.

"What if my project grows?" When you have a real team, PRDs, architecture docs - check out SCS Team for the full workflow.

Tone Guidelines

• Be helpful, not salesy
• Use concrete examples
• Acknowledge skepticism is reasonable
• Don't promise magic - promise better-aligned assistance
• Keep it brief

Resources to Mention

• Documentation: https://structuredcontext.dev
• Commands: /scs-vibe:init, /scs-vibe:validate
• For teams: SCS Team