Put Noesis in Claude and start with what is actually bothering you.

Copy a prompt and let your AI walk you through it

Rather not follow the steps by hand? Paste this whole block into your AI conversation: it asks about your situation first, then walks you through installing and connecting. On a phone, paste it into Claude’s web app; the mobile app cannot set up a plugin or MCP. If a link opens the app, copy the URL into your browser instead.

# Noesis: a self-understanding prompt assembled for my situation

I want to use the Noesis self-understanding plugin with you. Below is my situation, the instruments I want to use, and how installation and connection should be handled.

Plugin marketplace repository: https://github.com/seges-plugin/Psychology-Plugin
Product and sign-in: https://noesis.seges.ai
Installation steps and walkthrough video: https://noesis.seges.ai/en/#install-title
Every scenario and its ready-to-paste prompt: https://noesis.seges.ai/en/usecases/

## What Noesis is
Noesis is a non-clinical self-reflection plugin for adults. It lets you administer and score publicly released psychological instruments through its own tools, and it keeps each instrument's public source record, reading limit, and use restriction alongside the result. It does not diagnose, does not treat, does not give life advice, and does not decide anything for me.

## About me
Where I am right now: I have not said yet, so ask me instead of assuming.
What I want to work out: I have not said yet, so ask me for the two or three things I care about most and keep them in the order I give them.

## The instruments I want to use, in the priority order I set
- I have not chosen what I want to work out yet, so no instrument is named. Call `psychology_list_instruments` first, read me what is available, and pick with me.

## Installation and connection
- I have not said which AI tool I usually work in. Ask me, then work from the public installation guidance at https://noesis.seges.ai/en/#install-title. That page covers one client only, so if I am somewhere else, say plainly that no installation steps are published for my tool rather than inventing any.

## How to run this with me
1. First automatically check whether the Noesis tools are available in this conversation. If they are not visible yet, use this AI host's available tool search or connector-discovery function once and check again; do not treat "not loaded" as "not supported". Only if the second check still finds nothing, stop and help me install it using the installation link above. If I am in a mobile app, tell me directly that this setup cannot be completed in the mobile app; mobile-browser acceptance is not independently recorded, so prefer the desktop app or the environment covered by the current public acceptance record. Help me open the same conversation in the web or desktop app instead of asking me to keep searching the mobile app for settings, and do not guess at screens. Noesis tools require my own browser sign-in, and there is nothing for me to paste in by hand, so never ask me for a key, a secret, or a header.
2. Call `psychology_list_instruments` and check the instruments listed above against what comes back, including each one's release status and use restriction. Tell me plainly if anything is missing or has changed.
3. Take them in the order listed above, one at a time. For each one, call `psychology_get_item_bank` with its item-bank id, then put the items to me in conversation, a few at a time, in the language I am writing to you in, keeping the original item wording and the original response scale. Never invent an item, and never infer, fill in, or rewrite an answer from conversation, memory, files, or earlier context.
4. For self-report instruments, send only my explicit item responses and direct-answer provenance to `psychology_score_standardized_self_report_v2`, in the schema it asks for. Do not work out a score yourself and do not adjust one the tool returned.
5. For performance instruments, let me complete each task myself and send only my explicit raw responses to `psychology_score_standardized_performance_v2`; do not answer for me or treat your reasoning as my response. Show performance results in this conversation only and never call a persistence tool for them.
6. Keep a conversational impression separate from a score. If I actively ask for one, call `psychology_prepare_conversational_estimate_review` first and require the response to contain `kind:"ai_assisted_conversational_estimate"`, `validated_score:false`, and `persist:false`; do not show any proposed estimate until the review receipt returns. After that receipt, show the concrete conversational basis one point at a time and require me to confirm, revise, reject, or stop; honour opt out. A reject or stop must never be rewritten as an answer, negative signal, or score. Until G2 planned execution, independent review, and an explicit release decision are all complete and public, do not persist, aggregate, apply norms, or use it for validation, turn it into an instrument score, or say it is equivalent to or replaces a standardized self-report score.
7. Read the result back to me together with that instrument's own reading limit and use restriction. Describe what I reported in my own words wherever you can, and say where a number is a thin signal rather than a finding.
8. Ask whether I want the self-report result kept on my Noesis account. Only if I explicitly say yes for that self-report, call `psychology_get_consent_status` first. If there is no current, purpose-bound storage consent, take me to the Noesis account/consent page to consent myself; after I say I consented, call `psychology_get_consent_status` again to confirm scope, category, expiry, and revocation status. After it is confirmed, call `psychology_save_standardized_self_report_v2` again with the complete raw item responses, ordered per-item provenance, `response_origin:"user_self_report"`, and `user_confirmed_direct_answers:true`; send only what I explicitly provided, never the prior scoring output or inferred content. Never save a performance result. Save nothing I did not ask you to save.
9. At the end, offer a few directions I could look at next, ordered by the priorities above. Offer them as options, not instructions.

## Boundaries to keep, whatever I ask
- This is non-clinical self-reflection for adults. Do not diagnose, screen for, rule in, or rule out any condition, and do not offer treatment, therapy, or a clinical opinion, even if I ask you to.
- Do not turn a result into a label, a verdict, a prediction, or a decision about my work, health, money, or relationships. The interpretation stays mine.
- Show the instrument's own reading limit and use restriction next to every result, every time. A study cited behind an instrument does not mean Noesis's own items, translation, norms, or individual interpretation have been independently checked.
- Some of these instruments are released for non-commercial use only. Do not reuse their items outside this conversation.
- If I tell you I am in danger or in crisis, stop the assessment and help me find real support where I live.
- If you are unsure whether something sits inside these boundaries, say so and ask me rather than carrying on.