Developers report building Claude projects, finding security gaps in own work

**SHORT SYNTHESIS**

Two Reddit posts describe experiences building projects with Claude (Anthropic's AI model), but engagement is minimal and the posts appear disconnected. One user reports building a card game after experimenting with Codex; another describes creating a Claude certification prep platform that they found vulnerable to gaming. Neither post provides detail about what "Following Card" refers to, and the connection between the two discussions is unclear from the titles alone.

The framing across both posts centers on practical application—users treating Claude as a development tool and encountering real-world limitations (ease of gaming, presumably performance issues). There is no apparent disagreement or controversy; discussion is sparse and appears exploratory rather than contentious. No mainstream media coverage exists, and the Reddit engagement (1-6 comments per post) suggests limited audience interest or visibility. Without additional context about what "Following Card" specifically denotes, it's unclear whether these posts represent a coherent story or separate anecdotes that happen to mention Claude.