fix: Change 'never write insecure code' to 'generate secure code by default'

Changed absolute claim to more accurate statement:
- Before: 'never write insecure code' (too absolute, not provable)
- After: 'generate secure code by default' (accurate, acknowledges pattern)

Also updated READY_TO_POST.md to note we only tested with Claude Code.

All marketing materials now contain only verifiable claims.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

Author: silversurfer562

docs/marketing/READY_TO_POST.md

@@ -156,6 +156,7 @@ github.com/Smart-AI-Memory/empathy-framework
 ❌ "-75% linter errors" - Not precisely measured
 ❌ "Guaranteed ROI" - Too context-dependent
 ❌ "Proven adoption" - Too early
+❌ "Works with Copilot/Cursor" - Only tested with Claude Code
 
 ---
 

docs/marketing/TWITTER_THREAD_V3_9_3.md

@@ -10,7 +10,7 @@
 ## Tweet 1: Hook
 
 ```
-We just taught @AnthropicAI Claude how to never write insecure code.
+We taught @AnthropicAI Claude to generate secure code by default.
 
 The method: Put our coding standards in its project memory with real code examples.
 
@@ -146,7 +146,7 @@ To implement this pattern:
 1. Identify your top 5 coding violations
 2. Document with real code examples (not abstract rules)
 3. Show your actual implementation
-4. Add to .claude/ or .cursorrules
+4. Add to your AI tool's project context (.claude/ for Claude Code)
 
 pip install empathy-framework