Standalone fixes for all core skills. Added Standalone Quick Start guide to cali-product-shape-up so the LLM always runs Assumption Check before shaping. Input Detection sections now include inline text (verbal description) branches in plan-critique, codebase-critique, ux-critique, and execution-critique, so /skill:<name> <idea> works without a file path. Stelow artifact detection (.stelow///index.json) added to codebase-critique and ux-critique for appetite-aware gating. Workflow context detection added to testing-execution. index.json adopted as primary appetite source in interface-alternatives.
Português → English translation: all project files cleaned. Skills (ux-critique, codebase-critique, opportunity-mapping), references, extensions (.ts comments), ops scripts (install.sh, retired-skills.yaml), stages (stages.yaml), tests (test data), design docs (stelow-board-herdr.md), and CHANGELOG.md historical entries — all translated to English.
Changed
REVIEW_MODE default in cali-product-shape-up changed from "Product Spec + Interface + Scopes" to "Auto" for standalone (assumptions resolve internally instead of asking 5 questions).
Tool references standardized to CLI-agnostic pattern: ask_user_question → the ask tool throughout cali-product-shape-up.
cali-product-testing-execution and cali-product-testing-ai-code descriptions updated to mention standalone usage.
cali-product-scope-executor description now says "can be used standalone".
Fixed
Shape-up Input Detection omitted Assumption Check: standalone flow listed only "Parallel Recon → Shaping → Proposal output", skipping shape:15. Now lists all 5 steps.