- Home
- Skills
- Writing & Content
- Miantuan Pod
Works with the AI Agents you already use
Miantuan Pod
Miantuan Pod helps you recover an imaginary friend, dream companion, or childhood character you only half remember, and turn it into something you can see.
$2
Miantuan Pod
Example session with this skill installed
I had an imaginary friend when I was about twelve, but I can't remember much. Help me bring them back.
- Read your context and instructions
- Compiled the miantuan pod
Where we are: one anchor so far, a feeling (safe, a little mysterious), with night scenes beside the bed. There are no appearance details yet, so none will be invented.
Next question (one at a time, in your language):
Where did they usually appear?
- Beside the bed
- In dreams
- On the road home
- Somewhere in the sky
- Somewhere else
After a handful of short turns like this, the skill writes the dossier.
Reconstruction dossier (excerpt)
| Layer | Content |
|---|---|
| Remembered | About twelve years old. A calm, protective, slightly mysterious presence beside the bed at night. |
| Likely (inferred) | Closer to a night guardian than a playmate. Light seems part of their identity. |
| Uncertain | Name, face, exact outfit. These stay open instead of being filled in. |
| Visual direction | Soft white and grey palette, glowing edges, a silhouette that reads at small sizes. |
| Image prompt draft | Character concept art of a calm night-guardian companion. Preserve the remembered anchors. Soft glow, clean readable design, no text. |
Next steps
- Confirm or correct the dossier.
- Generate concept art and refine one detail at a time, keeping the approved ones.
- Optionally continue with
recall-to-petto prepare a pet handoff and, in Codex, the finished pet.
Connects securely to your tools. The creator never sees your data.
About this skill
Miantuan Pod helps you recover an imaginary friend, dream companion, or childhood character you only half remember, and turn it into something you can see.
What it does
- Interviews you one focused question at a time, with short options and an "I'm not sure" path, instead of a long questionnaire.
- Keeps remembered, inferred, uncertain, and proposed details apart, so a reconstruction is never passed off as a true memory.
- Produces a compact dossier and an image prompt, then iterates on concept art by changing only what you ask to change.
- Optionally continues into a Codex pet package and a public gallery entry, with only the final, approved material shared.
What is inside
| Skill | Role |
|---|---|
miantuan-pod | Entry skill. Reads what you want and loads the right skill below. |
recall-imaginary-friend | The recall interview, profiling, dossier, and concept art. |
recall-to-pet | Recall, concept art, pet hatching, and an optional gallery entry in one flow. |
structured-question | Asks one decision or preference question at a time. |
How it works
Say what you want, for example "help me remember my childhood imaginary friend". The entry skill picks recall only, recall and pet, or a single question, loads that skill, and the conversation continues in your language. If your agent has a question tool (opencode question, Claude Code AskUserQuestion, Codex request_user_input, dsh ask_user_question) it is used; otherwise the same question appears as a plain-text card.
Where it stops
- It reconstructs a faithful-feeling companion, not a verified memory, and slows down if the topic becomes distressing.
- Concept art needs an image generation tool or skill in your agent. Without one you get the approved image prompt to use elsewhere.
- Making a Codex pet needs Codex with its
hatch-petskill. In other agents the pack finishes at apet-handoff.mdfile you can take to Codex.
Source and updates: https://github.com/sci-m-wang/Miantuan-Pod (MIT license).
It reconstructs a faithful-feeling companion, not a verified memory. Concept art needs an image generation tool or skill in your agent; without one you get the approved image prompt. Hatching a Codex pet needs Codex with the hatch-pet skill; other agents stop at pet-handoff.md.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 10 days ago
Permissions
Read Files: the entry skill loads the sibling skills and any reference images you provide. Write Files: only to save an optional pet-handoff.md and concept art you ask for. Network Access: only when you ask for image generation or a GitHub gallery submission through your own tools; the skills themselves make no network calls.