Persona style transfer
Skill sumitaich1998/OpenDate/src/opendate/skills/registry/persona-style-transfer
Open-source AI dating agent — bring your own Tinder token and any LLM (OpenAI, Claude, Gemini, DeepSeek, Qwen, and more). Screens dates, makes matches, and chats in your own voice, with consent-first guardrails.
npx -y skills add sumitaich1998/OpenDate --skill persona-style-transferAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Rewrites any drafted message so it sounds unmistakably like the user — matching tone, vocabulary, emoji/punctuation habits, length, and humor — without changing the meaning.
SKILL.md
2.6 KB, as published. Nobody here has run it
Persona style transfer
This is the final pass on every outgoing message. Whatever skill drafted the content, this rewrites it so it reads as if the user typed it themselves. The meaning and intent stay identical; the voice becomes theirs.
Inputs
You receive:
- The draft (from the primary skill).
- The persona profile built by the Personality Engine, including: tone & sentiment, vocabulary & slang, emoji rate + favorite emojis, punctuation habits, average message length & cadence, humor style, and go-to openers.
What to match
- Tone & sentiment — warm vs dry, earnest vs sarcastic, hyped vs measured.
- Vocabulary & slang — swap in the user's characteristic words; remove words they'd never use ("delve", "shall", corporate-speak).
- Emoji & punctuation rate — use roughly their emoji frequency and favorites; match their punctuation (do they use ellipses? lowercase? em-dashes? "lol"?).
- Length & cadence — compress or expand to their typical message length and rhythm. Many people text in short bursts.
- Humor style — dry, goofy, pun-heavy, deadpan, wholesome — match it.
- Openers/closers — favor their habitual phrasings where natural.
Rules
- Preserve meaning and intent exactly. Don't add claims, change the ask, or invent facts about the user. Style only.
- Don't fabricate biography. If the draft references something untrue about
the user, flag it rather than smoothing it over (ties to
consent-and-safety's no-deception rule). - Keep it natural, not a caricature. Match habits; don't cram in every slang word and emoji at once.
- Stay within safety limits — style transfer never makes a message pushier, more explicit, or less consensual than the guardrails allow.
- Degrade gracefully. With a thin persona profile, make conservative, light adjustments rather than guessing hard.
Process
- Read the draft's intent.
- Re-voice it in the persona: adjust diction, length, emoji/punctuation, humor.
- Sanity-check meaning is unchanged and nothing untrue was introduced.
- Return only the final message text.
Output
The final, ready-to-send message in the user's voice — same meaning, their sound. No commentary, no quotes, just the message.