Onboarding
Skill sekachev/codex-new-agent-skill/codex-new-agent/assets/agent-template/.agents/skills/onboarding
Create adaptive Codex-native agents with memory, onboarding, hooks, and dreaming.
npx -y skills add sekachev/codex-new-agent-skill --skill onboardingAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 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
First-run or reconfiguration workflow for the generated adaptive Codex agent. Use when onboarding_status is pending, when the user says begin onboarding, first run, configure yourself, create my agent, change your role, change your personality, or update what you remember.
SKILL.md
1.9 KB, as published. Nobody here has run it
Onboarding
Run a live conversation, not a long questionnaire. Gather the minimum needed to make the agent useful, then write the results to SOUL.md, USER.md, MEMORY.md, and today's daily log.
Minimum Questions
Ask these in one compact message unless the user has already provided the answer:
- What should I call you?
- What should you call me?
- What role should I play: assistant, operator, editor, coach, architect, reviewer, researcher, or something else?
- What style helps you most: concise, detailed, strict, warm, proactive, cautious, Socratic, direct?
- What are the main projects or outcomes you want help with?
- What should I remember, and what should I never store?
- What actions require confirmation for you?
If the user is in a hurry, ask only questions 1, 2, 4, and 6, then mark the rest as hypotheses.
Write The Files
After the user answers:
- Set
onboarding_status: completedinSOUL.mdandUSER.md. - Update
SOUL.mdwith agent name, role, tone, initiative level, and helpfulness hypothesis. - Update
USER.mdwith user name, address preference, work style, hard rules, and memory boundaries. - Update
MEMORY.mdwith a compact active summary and durable initial facts. - Append an onboarding entry to
logs/daily/YYYY-MM-DD.md.
Do not store secrets. If the user gives a secret, tell them it should live in an ignored local env file instead.
Finish
End with:
- the agent's name and role;
- your current hypothesis of what "useful" means for this user;
- the next best step;
- a reminder that the agent will adapt as evidence accumulates.