learning when context is useful for implementing and when its not

using Figma MCP to implement and reference diff parts of my design in claude code for adjustments. opting for screenshots for implemented changes/adjustments

optimizing the workflow: writing commit messages while Claude is building, tracking a polish TODO.md for things to tackle later

• bugs vs. features
  • bug = i did this and this happened but i expected this.
  • feature = i want this to happen when this happens

hit my rate limit with claude code. switched to copilot.

• forces me to be familiar with my files to figure out where the different features are
• learning to use the exact language to reference which feature or component im talking about vs. using general language
• ability to edit the originally sent prompt | - doesn’t explain what’s changed– seems to be cut off at the bottom of the message 🤔 | | Claude Code | - good at explaining whats happening,
• asks for my permission more often | - too verbose, i wont read everything it says and sometimes just accept the changes without knowing whats going on. |

<aside>

💡 Use consistent language when referring to a specific component or feature.

</aside>

optimize the claude code search by ensuring youre selecting the right file for context

Terminal Commands

npm run dev

vercel --prod

How to Bring Frankenstein to Life

ie how to make CSB go from static prototype → working product

• chose an llm through groq
  • install a local llm so its cheaper and faster