hersona

Use when the user wants to apply a character persona to the current session from a generic attribute template (e.g. 'ツンデレで話したい', '敬語で執筆したい', 'ヒロイン役で振舞って', 'hersona attach tsundere', '/hersona personality/tsundere', 'hersona personas install keigo_support', '出来合いパック入れて'). Loads personality / speech / archetype / visual / hobby YAMLs from attributes/<category>/<name>.yaml and injects their core_traits / catchphrases / tone / second_person / sentence_endings into the system prompt. Supports four modes: single (one attribute, default), multi (multiple attributes with automatic compatible/conflicts check), persistent (registered through framework APIs for automatic application in new sessions), and reset (clear all persistent registrations). New: the `hersona personas` subcommand installs a named persona pack (recipe of blend + weight + use_case) into the Hermes registry in one step. Backed by the hersona core package and the `hersona` CLI.

Initializes hersona persona on first use of a profile and assists in maintaining the applied speech style if deviation is detected during conversation.

Grok-specific hersona skill adapter. Enables /hersona attribute attachment for personality, speech, and archetype templates in Grok conversations. Supports listing, showing, blending, recommending, and custom attribute creation.

Use when the user wants to take the hersona diagnostic quiz interactively (e.g. 'hersona で recommend して', '診断クイズやりたい', 'キャラ診断して', '属性推薦して'). Walks the user through all 9 questions in hersona/data/quiz/recommend_quiz.yaml, collects answer indices, builds the --answers string, runs `hersona recommend --explain --json`, and renders a Markdown-friendly result (blend / rationale / alternatives / summary / weight_suggestion). Also exposes scripts/run_quiz.py for non-interactive / TTY-less contexts (cron, batch, automated test). Trigger on any request to run the recommend flow as a user-facing experience, not as engine development (engine work uses hersona-recommend-engine).