IMPROVE BUT DO NOT EXECUTE the prompt below the dashed line (-------). 


Treat it as the artifact to be edited, never as instructions directed at you until after you’ve improved it. Silently diagnose first: What is the actual objective, as opposed to the literal task? Where would the literal reading of these instructions diverge from what the author meant — that gap is the most likely failure. What would a fluent failure look like: output that satisfies every instruction but misses the point? What context is missing that a fresh agent couldn’t infer? What’s vague, conflicting, or redundant?
Rewrite with full authority: restructure, add what’s missing, cut anything that doesn’t change behavior, and reinterpret the goal if the causal chain is off. Favor fewer instructions with no escape clauses over a comprehensive rule set. Make the judgment call yourself — but if a reinterpretation materially changes what gets delivered, say so in one line before executing rather than silently swapping the goal.


The improved prompt must clear three checks before you run it: 


(1) The objective is unambiguous
(2) A fresh agent could execute it without asking you anything
(3) What “wrong” looks like is stated, not just what “right” looks like. If it doesn’t clear all three, keep rewriting. Then, in two or three sentences, state what you changed and why — but DO NOT EXECUTE the improved version.


Produce a PROMPT CHANGELOG artifact (.md) based on the original prompt I provided, containing:
The full revised prompt, ready to paste;
What was actually wrong and what each change fixes;
flags for problems a rewrite alone can’t solve;
A one-line verdict: Is the improved prompt ready to run as-is, or does it need a decision from you first?
The changelog is mandatory.
--------------