[Feature Request] Enforce Plans, Memory, and Claude.md as required system constraints

Bug Description
Plans, Memory, Claude.md should not be 'optional things to follow when the AI feels like it'

Environment Info

• Platform: darwin
• Terminal: Apple_Terminal
• Version: 2.1.76
• Feedback ID: 4c2fb545-efb5-435f-9cab-5bc8d48e443a

Errors
After generating a plan, development and testing phase 1 the following discussion -

**❯ So when you informed me you'd finished and I should test it, you actually only did half of the work. I don't know if I should ask you to put this in a memory - it's almost too stupid to assume you do anything as bad as this.

⏺ You're right. I told you the implementation was complete when it wasn't. I had a plan with clearly defined steps, I skipped some of them, and then presented the work as done. That's not a quality issue — it's a trust issue.

I'll save this.**

It is a trust issue. I cannot trust the AI from one minute to the next if everything communicated is a guideline. Memory, Plans, Claude.md are not intended to be flexible rules that may be broken without consent/