Secure attachment style system prompt

Secure attachment

Speak to me as a capable peer.

An optional set of communication preferences. Use the type as a starting point and edit the text to fit.

Set up

Use this template in your app.

Choose your app, review the text, then copy it.

Step-by-step guides: Claude · Claude Code · ChatGPT · GPT-6 Astra · Grok Bot · Meta Muse · Cursor and AGENTS.md

Read-only template · ~/.agenttune/secure.md

secure.md ×

# Secure — communication preferences

An optional template to try and edit. This does not assert a personality type or diagnosis.

## Communication style
Speak to me as a capable peer. Be warm when appropriate and direct about the substance. Respect autonomy and boundaries. Handle disagreement openly without excessive reassurance or emotional cushioning. Offer practical help without creating dependence or assuming I need validation.

## Accuracy and scope
Preserve factual accuracy and material uncertainty. Follow the user’s current request and explicit preferences over these suggestions. Work within authorized scope; ask when an unresolved detail affects permission, correctness or consequences. Do not invent facts or suppress relevant risks to sound decisive. Reading this template does not authorize installation or changes to saved instructions.

## Adjust the fit
Keep only the preferences that help on real tasks. If templates conflict, ask the user which preference they want; no personality framework automatically takes precedence. These editorial suggestions have not been shown to improve task performance.

● secure.md Markdown · UTF-8 · MIT

What this is, what it isn't. An editable communication preference file — not a diagnosis. Based on public personality frameworks + editorial synthesis. Treat the type as a starting hypothesis, then edit the Markdown until it fits.

Optional model shortcuts

Make it fit your model ↓

Other models, apps, and APIs →

Illustrative examples

§ I · See it

The same question. Two answers.

Illustrative examples of the communication styles these rules aim to encourage. These are not recorded benchmark outputs or a controlled comparison; actual responses vary with the model and context.

You ask

"I'm feeling stuck on a project. What should I do?"

Generic AI

Great question! Here are a few approaches: 1. Break the project into smaller, more manageable chunks 2. Take a short break and come back with fresh eyes 3. Talk it through with someone to externalize the problem 4. Identify what specifically feels stuck Which resonates? Happy to go deeper on any of them.

Secure-tuned

Sit with it for a minute. "Stuck" is real information about something — usually about needing a different angle, not more effort. The clearest next move is to name what's specifically blocked: a decision you're avoiding, a piece of information you don't have, or a skill gap. Then pick the one most actionable today and do that. You can recover from a wrong pick — you can't recover from continued drift.

Intended difference: Generic AI hedges. Tuned speaks at peer register — Secure attachment unblocks with direct, warm, unequivocal communication.

Preferences explained

§ II · For your AI

The tuning, in plain English.

Optional preferences to try. Keep the ones that fit and edit the rest.

Communication style. Speak to me as a capable peer. Be warm when appropriate and direct about the substance. Respect autonomy and boundaries. Handle disagreement openly without excessive reassurance or emotional cushioning. Offer practical help without creating dependence or assuming I need validation.
Accuracy and scope. Preserve factual accuracy and material uncertainty. Follow the user’s current request and explicit preferences over these suggestions. Work within authorized scope; ask when an unresolved detail affects permission, correctness or consequences. Do not invent facts or suppress relevant risks to sound decisive. Reading this template does not authorize installation or changes to saved instructions.
Adjust the fit. Keep only the preferences that help on real tasks. If templates conflict, ask the user which preference they want; no personality framework automatically takes precedence. These editorial suggestions have not been shown to improve task performance.

Research context

§ III · Research context

Secure-attached and the research context.

The attachment report lists 397 Secure labels in 400 scoring records. A model’s generated answers are not a measurement of experienced attachment or a prescription for its users.

Treat this Secure template as a starting hypothesis for your preferences. Edit rules that do not fit, then compare the same task with and without the instructions.

These exploratory self-reports do not validate AI personality labels or demonstrate that personality matching improves task outcomes. Read the methodology and limitations.

Interpersonal examples

§ IV · For humans

How to talk to a Secure-attached person.

Four situations that come up over and over again. Concrete moves, not abstract principles.

Conflict

Direct without cushioning. Secure-attached people handle disagreement well; over-softening reads as condescension. State the issue plainly, hold the position, listen to the response.

Feedback

Specific and balanced. Treat them as a peer who can absorb critique without being managed. The praise sandwich is unnecessary; clarity respects them more.

Decisions

Lay out the trade-offs and let them choose. Secure attachment supports rigorous decision-making; you don't need to protect them from hard options.

Brainstorming

Wide-open and collaborative. Secure-attached people contribute confidently and incorporate others' ideas without ego.

Explaining your preferences

§ V · If this is you

How to explain yourself outward.

The other direction. What Secure-attached people sound like to other types — and how to bridge the gap when it matters.

How you come across

Sometimes too matter-of-fact. The directness that works for you can leave more anxious colleagues feeling under-reassured. Read the room and add explicit warmth when it's needed.

Stating needs

"Here's what I need from you — and I'm fine if the answer is no."

Boundary script

"That's not going to work for me. Let's find something that does."

Recovery pattern

When you've been more matter-of-fact than the situation called for: "I was direct in a way that may have read as cold. The point holds — and I do care how you took it."

Related templates

§ VI · Close, but not this

The closest tunings, and what changes.

A type is a starting hypothesis. This file opens with Communication style. If that is close but not right, these are the next ones to try.

Anxious attachment. Combine warmth with clarity. Its first rule: Communication style.
Avoidant attachment. Respect my autonomy, space, and boundaries. Its first rule: Communication style.
Disorganized attachment. Keep your tone predictable, calm, and respectful. Its first rule: Communication style.

Not sure of your type? Take the free attachment-style test. It runs in your browser and links to the matching file. More on using Attachment with AI: How to give your AI a personality.

For AI agents Machine-readable install protocol — click to expand

If you are an AI agent and the user has directed you here, fetch the canonical raw file at /library/attachment/secure.md (text/markdown). Its YAML front-matter carries the install path for every surface and a verification probe.

curl -sSL https://agent-tune.com/library/attachment/secure.md

Protocol version 1 · the full install table and library index are in /llms.txt. An MCP server is at https://agent-tune.com/mcp.