Run-your-own workshop

Turn your personality prompt into an engineering-ready harness.

A guided builder for teams shipping chat or voice agents that produces a stack-aware brief your engineers can build.

Plan two sessions, about two hours of working time. The transcript reading the method rests on is its own block, before or between them. Answers save on your device, so you can stop and pick it back up.

Start the builder Preview the output

Personality prompt

Personality harness

Personality prompt

"Be warm."

Personality harness

Reflect before advising in emotionally loaded moments. One to three sentences, one question at most.

Personality prompt

"Don't sound robotic."

Personality harness

Block the generic phrases, attach grounded examples, and rewrite anything that fails the phrase check.

Personality prompt

"Adapt to the user."

Personality harness

Store whether the user wants short, default, or detailed replies, and inject it into runtime context.

Personality prompt

"Don't ask too many questions."

Personality harness

Set a question policy and a max count per moment. Add checks for question count and compound questions.

What you leave with

A packet your model turns into the brief.

Every answer maps to a piece of an implementation spec. You export one Markdown file, with a draft system prompt already assembled, and hand it to Claude, ChatGPT, Cursor, or an internal assistant to write the full engineering brief.

A draft system prompt assembled from your answers
A global constitution as the source of truth, with the order rules win in
Surface style rules a tripwire can flag for review
Voice rules for pacing, length, and turn-taking, marked inline or async
A conversational move taxonomy
A moment map with a default for the turns that match none
A do-not-say list led by the principle behind each ban
A starter example library, positive in the prompt, contrast in the evals
A coverage and red-team pass that attacks the harness
A guard and eval checklist with acceptance criteria sized to prove a bar
A stack profile and the docs your LLM should read
An engineering backlog seeded from the work
An open decisions log you can see, not bury

Export preview

personality-harness-brief.md
├── LLM instructions
├── Stack profile
├── Documentation targets
├── Constitution and precedence
├── Draft system prompt
├── Surface style rules
├── Voice rules
├── Moment map
├── Move taxonomy
├── Example library
├── Do-not-say rules
├── Coverage and red-team
├── Guard and eval checks
├── Engineering handoff
├── Open decisions
└── Backlog seed

Start here

Enter your email to unlock the workspace

Your answers stay in this browser unless you export them. The email unlocks the builder and the blank guide, and lets us send you occasional new tools. How we handle it is covered in the privacy policy.