Replit Agent adds 61 games to a family game vault
Andrew and Kevin Blumson asked Replit Agent to grow their family game vault from 40 to 101 titles, with no repeated games, then shared play-testing caveats.
x.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Expand our family mini-game vault from 40 to 101 distinct games. Do not duplicate a game even if its title differs; test that levels run, then flag content that needs a person to judge for fun and age suitability.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- Andrew Blumson (@Andrew_Blumson) and Kevin Blumson (@KevinBlumson)
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The reported game collection grew by 61 games, from 40 to 101. The builders say technical playability did not establish that every game was enjoyable.
- Human involvement
- The builders defined the collection constraint and distinguish functional level testing from human judgments about fun. They invited community play-testing.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- The builders set a target to expand the game vault from 40 titles to 101.
- They gave Replit Agent one explicit constraint: do not repeat games; genres could repeat.
- They used automated play-throughs to check that levels worked, then asked people to judge whether the games were fun.
Andrew Blumson reported that Replit Agent autonomously added 61 games to the family’s mini-games app. A related post from co-builder Kevin Blumson explains the target: the vault had 40 games and they asked the agent to bring it to 101, with one explicit rule—no repeated games. The rest of the selection was left to Replit Agent.
The pair also described how they checked the generated content. Kevin says Replit Agent play-tested all 200 levels in games with that many levels, and around 400 levels for endless procedural games. He immediately points out the limit: a level that technically works is not necessarily fun. The family invited people to try the games and share feedback, so the reported result is a larger playable collection, not proof that every game was enjoyable or suitable for every child.
This is a useful way to expand a small game library: define the quantity and one hard constraint, let the agent fill the catalogue, then test functional coverage separately from play quality. The source does not publish the full prompt or a list of all 61 additions. The suggested prompt below is editorial, and a human should still check age suitability, accessibility and whether games repeat under different names.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: X · Andrew Blumson (@Andrew_Blumson) and Kevin Blumson (@KevinBlumson)
Frequently asked questions
How many games did Replit Agent add?
Andrew Blumson reported 61 additions, bringing the family game vault from 40 to 101 titles according to co-builder Kevin Blumson.
What rule did the builders give the agent?
Kevin says the only rule was not to repeat games; repeated genres were allowed.
Did automated play-testing prove the games were fun?
No. Kevin says technical playability did not guarantee fun and asked people to try the games and give feedback.
More details
About this account
Replit Agent chose and added the games under a target count and no-duplicate rule, and the builders say it play-tested game levels.
Reviewed archived direct X capture from 2026-10-15 (HTTP 200), including Andrew’s post and Kevin’s related reply, authors and displayed date; current reader returned 403. No exact prompt or full game list was shared.
Credit: Andrew Blumson (@Andrew_Blumson) and Kevin Blumson (@KevinBlumson) — X
Source published
Added to Agents In Use
Source checked
Tools and topics
Replit Agent
family games · game testing · content generation
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.