Repository navigation
Have AI analyze submissions async to detect spam #101
Description
Activity
Would need to double check, but Anthropic very likely happy to fund API keys for this purpose :)
Closing this out as speculative. The moderation approach that settled in post-launch leans deliberately lightweight (see moderation policy), relying on upstream registries and manual triage rather than an AI pre-filter. The related #98 is also closed. Easy to reopen if someone wants to drive a concrete design.
As an anti-spam measure, we could have an LLM reviewing submissions and flagging abnormalities, perhaps with automated suggestions for the blacklists in #98.