voice
/voice
Prowadzi adaptacyjny wywiad o stylu pisania, by zbudować wielowymiarowy profil głosu użytkownika.
- Autor
- zircote ↗
- Repozytorium
- human-voice ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 20.07.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 1
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 20.07.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Prowadzi adaptacyjny wywiad o stylu pisania, by zbudować wielowymiarowy profil głosu użytkownika.
Oryginalny opis (EN)
This skill should be used when the user asks to start a voice interview, elicit writing style, build a voice profile, analyze writing voice, capture writing style, run voice elicitation, understand my writing style, profile my voice, or needs to conduct an adaptive interview that produces a multi-dimensional voice coordinate profile.
Jak działa
- Architecture Overview
- Interview Modules
- Starting an Interview
- Resuming an Interview
- Profile Output
- Data Directory
- Session Storage
Podgląd SKILL.md
---
name: voice
description: This skill should be used when the user asks to start a voice interview, elicit writing style, build a voice profile, analyze writing voice, capture writing style, run voice elicitation, understand my writing style, profile my voice, or needs to conduct an adaptive interview that produces a multi-dimensional voice coordinate profile.
allowed-tools:
- Read
- Write
- Bash
- Glob
- Grep
- AskUserQuestion
- Agent
---
# Voice — Voice Elicitation Interview Engine
Voice is an adaptive voice elicitation interview engine that builds a comprehensive written-voice profile through a structured conversational interview. The engine administers 60-80 questions across 12 thematic modules, adapting its path based on the participant's responses and detected writer type.
## Architecture Overview
- **Question Bank**: JSON question files organized by module in `question-bank/modules/`, validated against `question-bank/schemas/question.schema.json`
- **Branching Logic**: Conditional paths in `question-bank/branching/` that route participants through writer-type-specific question sequences
- **NLP Pipeline**: Text analysis stages in `nlp/` that extract stylistic signals from open-ended and writing-sample responses
- **Scoring Engine**: Dimension scoring in `scoring/` that maps responses to voice profile dimensions with weighted aggregation
- **Agents**: Specialized conductor and analyst agents in `agents/` that orchestrate the interview flow
## Interview Modules
The interview spans 12 modules plus a screening/demographic preamble (SD). Modules cover dimensions such as formality and register, emotional tone, sentence structure, vocabulary preferences, rhetorical patterns, audience awareness, and stylistic identity. Each module contains 4-8 questions with a mix of question types (likert, forced_choice, scenario, open_ended, writing_sample, calibration, etc.) to triangulate self-reported preferences against observed behavior.
## Starting an Interview
1. Generate a UUID session ID.
2. Create the session directory at `~/.human-voice/sessions/{session_id}/`.
3. Initialize `state.json` with status `init`, empty response arrays, and module progress tracking.
4. Load the question bank from the project's `question-bank/modules/` directory.
5. Begin with the screening/demographic module (SD) to establish baseline writer type and branching path.
6. Progress through modules in sequence, applying branching rules and format rotation to maintain engagement.
## Resuming an Interview
1. List session directories under `~/.human-voice/sessions/`.
2. Identify sessions with status `paused` or `in_progress` in their `state.json`.
3. Load the selected session's state, including the last answered question, current module, and accumulated responses.
4. Resume from the next unanswered question, preserving all prior scoring and branching decisions.
## Profile Output
Voice produces a dual-output voice profile stored as `profile.json` in the session directory:
- **Self-Report Layer**: Direct preference scores derived from likert, forced_choice, and semantic_differential responses. These capture what the participant believes about their own voice.
- **Computational Analysis Layer**: Scores derived from NLP analysis of open_ended, writing_sample, and process_narration responses. These capture what the participant actually does in practice.
- **Calibration**: Calibration questions cross-reference self-report against behavioral signals to produce a calibration delta — the gap between perceived and actual style. High deltas flag dimensions where the participant's self-image diverges from their practice.
The final profile includes per-dimension scores, a composite writer-type classification, distinctive features, and calibration annotations.
## Data Directory
**All profiles, sessions, and config live under `~/.human-voice/`.** This is the single canonical location — the plugin does not honour plugin-runtime data directories, ensuring consistent profile access across multiple Claude accounts.
## Session Storage
All session data is stored under `~/.human-voice/sessions/{session_id}/`:
- `state.json` — Session state: status, current module, current question index, branching path, timestamps
- `responses.json` — Raw responses keyed by question_id with timestamps
- `scores.json` — Running dimension scores updated after each response
- `profile.json` — Final assembled voice profile (written on completion)
- `writing_samples/` — Collected writing samples for NLP analysis
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · zircote/human-voice | 90 | MIT |
| chatcut-inc/agent-plugin | 100 | Apache-2.0 |
| chatcut-inc/agent-plugin | 100 | Apache-2.0 |
| crisandrews/clawcode | 100 | MIT |
| simota/agent-skills | 100 | MIT |
| conversiontools/agent-skills | 100 | MIT |
| arach/lattices | 100 | MIT |
| guaardvark/guaardvark | 100 | MIT |
| lauradfarbo/chief-of-staff | 100 | MIT |
Migawka bazy z 13.11.2026 · wersja paczki e9c4123326ce (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 2 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100voice-calibration
Uczy się osobistego stylu pisania i mówienia przez interaktywne zadania, a potem tworzy profil głosu do wielokrotnego użycia. - 100profile-interview
Przeprowadza wywiad pytanie po pytaniu, by zbudować profil głosu i gustu użytkownika, z którego inne skille piszą teksty jego głosem. - 100voice-interview
Przeprowadza wywiad i zapisuje profil stylu pisania: rytm, otwarcia, zakończenia i zwroty, do pliku core/voice-profile.yml. - 100setup-writing-style
Uczy się stylu pisania użytkownika z jego wysłanych wiadomości i dokumentów i buduje profil głosu zapisany jako skill my-writing-style, by szkice nie brzmiały generycznie. - 100writing-style
Konfiguruje profil głosu na dwóch-trzech próbkach tekstu i stosuje go do długich treści, aby pisanie brzmiało spójnie. - 100blog-style
Uczy się stylu autora z 5 do 10 wpisów i tworzy profil głosu, potrzebny do pisania w tym samym tonie.