Make /vscode-pet a Tamagotchi for AI agents 🐣
- Dominant language
- TypeScript
- Stars
- 193k
- Forks
- 42.4k
- PR merge metrics
- PR metrics pending
Description
I really like the idea of /vscode-pet. I think it could be even more interesting if the pet behaved like a Tamagotchi for the AI agents.
The pet could have simple needs and states that are connected to agent activity:
* Tokens could represent food for the pet.
* Running agents could give the pet energy.
* When all agents are idle, the pet could go to sleep.
* Agent activity could give the pet experience.
* Over time, the pet could grow and evolve.
* The pet could have simple states such as hungry, happy, tired, or sleeping.
For example:
/vscode-pet feed
could feed the pet based on the tokens being consumed by the agents.
/vscode-pet play
could trigger a simple interaction with the pet.
/vscode-pet status
could show the pet’s current state and the activity of the agents.
When agents are actively working, the pet could be awake. When all agents finish, the pet could go to sleep.
The fun part is connecting the pet to token usage. Instead of tokens being just a number in a usage dashboard, they could become something tangible: the pet is effectively being fed by AI usage.
For example, if the pet becomes hungry, the user could interact with it by using their AI agents and consuming tokens. As the agents work and consume tokens, the pet gets fed, becomes happier, and eventually evolves.
This could also create a natural engagement loop around Copilot usage:
* Pet is hungry → user uses the agents → tokens are consumed → pet gets fed.
* Agents are actively working → pet is awake and energetic.
* Agents are idle → pet eventually goes to sleep.
* Continued usage → pet gains experience and evolves.
The pet wouldn’t need to artificially consume tokens or encourage unnecessary usage. It could simply reflect the user’s existing AI usage in a playful way.
This could make token usage more tangible while also creating a fun reason for users to keep coming back to their pet.
Contributor guide
Assessment
This issue has not been assessed yet.