Read catalog.json
Fetch catalog.json. It carries generated, count, categories[] and a skills[] array whose entries mirror each meta.json. Choose by id, category or tags.
Machine-consumable knowledge base
Every project's method, written down as a reusable skill.
The working methods behind Dr. Sohil Momin's projects — the offline-first single-file web app, Android WebView packaging, the portable EXE launcher, the shared release workflow, the landing-page conventions, the open-data pipeline, the LAN server architecture and optical file transfer — captured so another AI, or a future project, can reproduce the work instead of starting over.
The catalogue
Grouped by category. Filter with the chips, or search across name, description, tag and source repo. Every card links to its SKILL.md, its meta.json and the projects it was derived from.
/
No skills match that filter. Try a different word, or choose All.
How an AI agent consumes this repo
The catalogue is the single source of truth. Read it first, then the one skill you need.
Read
Read the SKILL.md
Each skill is YAML front-matter (name, description) plus When to use, Method, Key files / entry points, Gotchas and Source projects — specific enough to reproduce the work.
Check freshness
Compare meta.json's source_snapshot (repo → commit SHA) against the source repo's HEAD. A weekly workflow keeps the SHAs and catalog.json current, but never rewrites the prose.
A compact pointer file for agents lives at llms.txt.