52 prompts indexed, 38 production-ready, used daily by the operator

The claim

The catalog is a working tool, not a deck slide: 52 prompts indexed, 38 promoted to a tested-and-ready status, exercised by the operator across active engineering sessions every working day.

The evidence

The repository carries 52 prompt artefact directories, each with a typed metadata file, a templated body, and most with eval cases. 38 carry a ready lifecycle flag (gated promotion, not self-declared); 2 sit at tested; 11 remain draft. The catalog has shipped real downstream work: it generated the CRUMB flashcard extractor, the brand-asset generator, and 32 governance prompts that run the operator’s daily project oversight.

Because the operator dog-foods the tool against their own daily engineering workload, every prompt in ready carries a signed eval and at least one shipped use against a paying or pre-paying surface in the operator’s portfolio. There is no synthetic activity in the usage log.

The comparable

Most pre-seed developer tools at this stage carry zero shipped artefacts of their own; EVA carries 38. For frame: the median YC W23 batch dev-tools company entered Demo Day with a v0 internal demo and a waitlist. EVA enters with a working library that has produced 38 production-grade outputs.

What we ask for

A check sized to the next milestone — a hosted multi-user index — not a redo of what already works. The thesis card frames the wedge; this card is the proof that the wedge has been cut.

xychart-beta
  title "Catalog status across 52 prompts"
  x-axis [ready, tested, draft]
  y-axis "Prompt count" 0 --> 50
  bar [38, 2, 12]