mirror of
https://github.com/garrytan/gstack.git
synced 2026-11-23 00:21:14 +01:00
In Claude Code, every gstack skill now runs its start command and loads the shared preamble while the skill loads, so the model starts with their output instead of an instruction it can skip. Why: in test runs, loaded skills' steps were followed in only 45% to 62% of sessions, and /autoplan's start command never ran in 5 of 10. Nothing to do: other hosts keep today's instructions, and if Claude Code can't run the step, the skill says so and the model runs it as before. Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com> Review-Finding: p1/specialist-batched#1 reader=security Review-Finding: p1/specialist-batched#3 reader=testing Review-Finding: p1/red-team#1 reader=red-team Review-Finding: p1/red-team#2 reader=red-team