A prompt injection solution that seems to benchmark better than any other approach out there, while not using hard-coded filters or a lightweight LLM which adds latency.

https://archigraph.ai/ An architecture-level agentic IDE.

I am working on ways to measure cognitive performance throughout the day and the impact of supplements or activities that supposedly help

Nocode transform: From 30 photos --> FPS video game.

Example : https://shorturl.at/We3dH

Building a tool to make test reports useful - https://gaffer.sh

Building a privacy first subscription tracker. https://paylog.space

Value-study-practice-o-matic is the best name I've come up so far. It's just an automated way of practicing rendering when painting.

I'm building a zork-like dungeon explorer for vibe coded projects. Ok, the zork interface is not that important, but it adds an extra layer of fun, and does reflect the reality of how I dig through a codebase to understand it. You start at the entry point and start exploring each code path to build a map of what is going on, taking notes as you go, and using tools if you're lucky to get a sense of the overall structure. You can also go up and down a level of abstraction like going up and down a dungeon.

It incorporates also complaints from a static analyzer for Python and Javascript that detects 90+ vibe slop anti-patterns using mostly ASTs, and in some cases AST + small language models. The complaints give the local class and methods a sense of how much pain they are in, so I give the code a sense of its own emotional state.

I also build data flow schematics of the entire system so I can visualize the project as a wire diagram, which is very helpful to quickly see what is going on.

Honestly, I thought that if I had an AI knowledge base instead of code, anything would work, so I created this OpenID Connect OP (IdP) that processes data in real time according to specifications using AI. Have fun! https://github.com/snakaya/GenAI-OIDC-IdP