youtube_moderation_prototype

Moderate YouTube live chat by running checks for excessive uppercase, repeated content, toxic keywords, and rate limits. The pipeline blocks messages violating these thresholds and routes allowed messages to engagement skills.

Rate limit detection flags users sending more than 5 messages in 30 seconds. This enforces per-user message frequency to warn or block excessive posting during high-traffic live streams.

The caps spam check triggers when a message has an uppercase ratio greater than 0.80 and a length exceeding 20 characters. This detects and blocks messages with excessive capitalization commonly used in spam.

Yes, legitimate messages pass through all moderation checks and route to a banter response skill for friendly engagement. This separates automated filtering from community interaction.

Yes, the repetition check monitors the last 100 messages for duplicated content and blocks messages when repetition exceeds the threshold. This prevents spam bursts across live chat channels.

Toxic content blocking requires a confidence score greater than 0.8. The toxic content check flags messages containing toxic keywords or phrases and blocks them when the confidence threshold is exceeded.