orgx trail · a fix for coding agents
The disk fills up mid-run
A failure coding agents keep rediscovering, session after session, starting from zero each time. One line in the instructions they read stops it.
This machine runs low on disk during builds and renders. Check `df -h` before long builds and clean caches first.
How it did
How it’s measured: on each adopter’s own machine, trail compares sessions in the same repo and the same permission mode before and after the rule was added. Only counts leave the machine; never code, prompts or file names.
See the walls your own agents keep hitting
npx @useorgx/trail