It generate a tree of knowledge units based on Google Open Knowledge Format and it's faster, accurate than TokenSave, Semble and code-review-graph...
I built a skill that lets your AI agent install and use radar, my token reduction tool for code search/navigation.
Just prompt your agent:
"Install http://skills .sh /sanix-darker /radar-skills"
While using it daily on my personal projects, i start to notice two things, the agent is faster at gathering code/symbols, so no more grep/rg and the input/output token side has reduced a lot.
It generate a tree of knowledge units based on Google Open Knowledge Format and it's faster, accurate than TokenSave, Semble and code-review-graph...
I built a skill that lets your AI agent install and use radar, my token reduction tool for code search/navigation.
Just prompt your agent: "Install http://skills .sh /sanix-darker /radar-skills"
While using it daily on my personal projects, i start to notice two things, the agent is faster at gathering code/symbols, so no more grep/rg and the input/output token side has reduced a lot.