# Changelog — sample-petcare-consultative

## v0 — 2026-06-15
**Origin**: Sanitized reference sample derived from a production pet-education/consultative agent. Not deployed.
**Template**: principles
**Flavor**: consultative
**Deploy**: none — teaching sample, never written to any DB.
**Notes**: PII stripped (real brand, founder/instructor names, social handle, all real URLs, WhatsApp `wa.me` links, Google Drive PDF links, team/staff rosters, interview-booking specifics, Slack channels, location specifics, real influencer/agent UUIDs). Curriculum specifics (hours, month ranges) genericized to "flexible professional curriculum". The reasoning-leak guard (Principio 1 — disfraz humano + sin razonamiento visible) was added to `sdk/system.md`. Source scripted prompt was re-expressed in the `principles` shape while preserving the consultative motion (5+ turn discovery, mid-conversation link drop, same-day DM close) and the multi-dimensional qualification architecture (occupational fit / role clarification / modality fit).