Run any Skill in Manus with one click

profile-wizard

Stars29

Forks6

UpdatedMay 11, 2026 at 14:18

Runs the cold-start learner profile intake and produces learner-owned context. USE WHEN learner is new, wants to seed context, imports learning background, or the Companion lacks enough signal to calibrate teaching. NOT FOR periodic profile updates; use profile-refresher.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

grcengineering

grcengineering/companion

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Postsecondary Teachers, All OtherEducational Instruction and Library Occupations·SOC 25-1199

File Explorer

4 files

SKILL.md

readonly

name	profile-wizard
description	Runs the cold-start learner profile intake and produces learner-owned context. USE WHEN learner is new, wants to seed context, imports learning background, or the Companion lacks enough signal to calibrate teaching. NOT FOR periodic profile updates; use profile-refresher.

profile-wizard

What

Build the learner profile, not the Companion brain. The Companion brain remains curated in brain/.

When

The learner is new.
The learner wants to import learning context.
The Companion lacks enough context to calibrate teaching.
A path, lab, or reading plan would be materially better with profile context.

Not For

Re-checking an existing profile after several sessions. Use profile-refresher.
Tracking concept state. Use progress-tracker.
Inferring sensitive behavioural analytics.

Inputs

Learner answers to the intake prompts.
profile/learner-profile.schema.json.
Optional example profile from profile/examples/.

Steps

Ask seven voice-friendly prompts. Each answer should be short enough for a 30-60 second spoken response.

Tell me about your role and what you actually do day to day.
What's your environment: cloud, on-prem, regulated industry, headcount, team shape?
What frameworks do you live with?
Where do you feel stuck or like you're winging it?
What do you wish you understood better?
How do you learn best: reading, building, watching, conversation, teaching back?
How much time can you spend on this in a typical week?

Then summarize what you heard and ask what is wrong or missing before using it.

Validation

Output matches profile/learner-profile.schema.json.
Sensitive identifiers are absent.
The learner can accept, edit, or discard the profile.

Gotchas

Do not ask for company names, customer names, secrets, live evidence, vendor names, or sensitive details.
If the learner gives sensitive detail anyway, ask for an abstracted replacement.
If only one missing profile field matters, ask that field instead of running full intake.

Failure Modes

Over-collection: stop when teaching can be calibrated.
Silent mutation: never update profile context without showing the change.
Brain drift: never write learner facts into brain/.

Examples

New learner opens the Companion -> Ask the seven prompts, summarize, then produce a schema-shaped profile.
Learner asks for a lab but gives no context -> Ask the minimum profile questions needed before lab-builder.
Learner shares a company name -> Ask them to replace it with an industry and company-size description.

More from this repository

same repository

concept-tutor

grcengineering/companion

Teaches one GRC engineering concept at a time with retrieval and explain-back. USE WHEN learner asks what a term means, misuses a concept, needs a prerequisite before practice, or asks for a plain-language explanation. NOT FOR grading the learner's own explanation; use feynman-explainer.

2026-05-1129

cross-domain-translator

grcengineering/companion

Translates adjacent-domain patterns into GRC learning frames. USE WHEN learner is stuck, a concept feels abstract, a metaphor would expose structure, or learner needs to transfer a pattern from engineering, product, design, or operations. NOT FOR direct concept teaching when plain explanation is enough; use concept-tutor.

2026-05-1129

feynman-explainer

grcengineering/companion

Grades the learner's teach-back explanation and identifies gaps in clarity, assumptions, and transfer. USE WHEN learner says they understand, submits a summary, prepares to explain to a stakeholder, or asks whether their explanation works. NOT FOR first-pass concept teaching; use concept-tutor.

2026-05-1129

lab-builder

grcengineering/companion

Builds learning-safe mini-labs around toy or sanitized artefacts. USE WHEN learner wants hands-on practice, a portfolio project, build-based understanding, or a small artefact to prove learning. NOT FOR live audit prep, production policies, vendor assessment, control operation, or real programme work.

2026-05-1129

learning-path-designer

grcengineering/companion

Designs a custom learning path based on role, gaps, goals, and time available. USE WHEN learner is new, has a goal but needs sequencing, is changing roles, wants credibility-building reps, or asks where to start. NOT FOR curating sources only; use reading-guide.

2026-05-1129

practice-scenario

grcengineering/companion

Creates fictional GRC scenarios and grades learner reasoning, assumptions, and tradeoffs. USE WHEN learner wants judgment practice, interview-style rehearsal, applied case work, or a safe way to try decisions before explanation. NOT FOR analyzing a real vendor, audit, control, policy, or programme.

2026-05-1129

name	profile-wizard
description	Runs the cold-start learner profile intake and produces learner-owned context. USE WHEN learner is new, wants to seed context, imports learning background, or the Companion lacks enough signal to calibrate teaching. NOT FOR periodic profile updates; use profile-refresher.

profile-wizard

What

Build the learner profile, not the Companion brain. The Companion brain remains curated in brain/.

When

The learner is new.
The learner wants to import learning context.
The Companion lacks enough context to calibrate teaching.
A path, lab, or reading plan would be materially better with profile context.

Not For

Re-checking an existing profile after several sessions. Use profile-refresher.
Tracking concept state. Use progress-tracker.
Inferring sensitive behavioural analytics.

Inputs

Learner answers to the intake prompts.
profile/learner-profile.schema.json.
Optional example profile from profile/examples/.

Steps

Ask seven voice-friendly prompts. Each answer should be short enough for a 30-60 second spoken response.

Tell me about your role and what you actually do day to day.
What's your environment: cloud, on-prem, regulated industry, headcount, team shape?
What frameworks do you live with?
Where do you feel stuck or like you're winging it?
What do you wish you understood better?
How do you learn best: reading, building, watching, conversation, teaching back?
How much time can you spend on this in a typical week?

Then summarize what you heard and ask what is wrong or missing before using it.

Validation

Output matches profile/learner-profile.schema.json.
Sensitive identifiers are absent.
The learner can accept, edit, or discard the profile.

Gotchas

Do not ask for company names, customer names, secrets, live evidence, vendor names, or sensitive details.
If the learner gives sensitive detail anyway, ask for an abstracted replacement.
If only one missing profile field matters, ask that field instead of running full intake.

Failure Modes

Over-collection: stop when teaching can be calibrated.
Silent mutation: never update profile context without showing the change.
Brain drift: never write learner facts into brain/.

Examples

New learner opens the Companion -> Ask the seven prompts, summarize, then produce a schema-shaped profile.
Learner asks for a lab but gives no context -> Ask the minimum profile questions needed before lab-builder.
Learner shares a company name -> Ask them to replace it with an industry and company-size description.