fix(skill-creator): keep scanning trigger events - #1766

Flagging a collision risk before this grows further: skills/skill-creator/scripts/run_eval.py trigger detection is being edited by four open PRs from the same base (e58c70bea3):

• fix(skill-creator): scan full turn for skill triggers instead of bailing on first non-Skill tool (#1721) #1757 (toprakeker) — rewrites this exact logic as a TriggerDetector state machine for skill-creator: trigger detection reports 0% recall for every skill #1721
• fix(skill-creator): make the trigger-eval harness measure skill triggering #1755 (giovannibeltrame) — reworks the trigger-eval harness
• fix(skill-creator): isolate trigger evals and handle Windows and runtime failures #1298 (MartinCajiao) — broader fix for the 0% recall issue, still actively updated
• this PR — minimal patch for the same bail-on-first-non-Skill-tool bug

Your patch at L138+ fixes the same short-circuit behavior #1757 addresses with a more complete state machine, and the hunks will conflict. Suggest coordinating with #1757's author (either fold this into it or rebase on top once one of them merges). Thanks!

Closing this smaller overlapping patch in favor of #1757, which addresses the same trigger-scanning defect with the complete state-machine implementation and dedicated tests. Thanks for pointing out the collision.