meta prompt

The Conversation Audit & Codify

Building the Brand Writing Team, Live

Models
claude-sonnet-4-6
Updated

Use it for

Building the Brand Writing Team, Live

The prompt

I've been iterating on [ARTIFACT — e.g., a skill, a prompt, a piece of code, a document, a process doc] over the course of our recent conversations and edits. I want you to close the loop and bake every improvement back into the source of truth at the root-cause level.

Do the following:

1. Audit our recent conversation(s) on [ARTIFACT]. Identify every moment where one of these happened:
   - I made a decision about how it should work
   - We found and fixed an error or inconsistency
   - I added a new requirement, constraint, or quality bar
   - I clarified something you had misunderstood
   - You produced a fix or workaround in the moment but didn't update the source file

2. For each item, figure out where it belongs in [ARTIFACT]. Not a patch on top of the symptom — the root-cause location.

3. Update [ARTIFACT] with all of those changes, reflecting each one at the appropriate level of abstraction.

4. Show me a concise changelog: what you changed, where, and which conversation moment it came from. Keep the changelog in chronological order.

Do not skip items you think are minor. Small, repeated frictions compound; that's exactly what this audit is for.

Worked example

No worked example recorded yet.