Skip to content

Improve keyboard-mash / gibberish detection #57

Description

@Geph

Summary

Repeated-character spam is blocked (6+ identical chars), but high-entropy keyboard mash still passes moderation.

Examples that currently slip through

  • NO ICE JBNJKJHBVBNJKLKDJHVCSVBNMKALSDCJHGVABNJKSZXCVJHBBNXJCHBVBDNJFKJDHEGBDNRJFKOIDUHGEBFNMKLDEIHGBN
  • asdfghjkl, qwertyuiop
  • Short junk under the 8-letter diversity threshold

Suggested approach

  • Score qwerty-adjacent / common mash patterns
  • Reject “meaningful prefix + junk suffix”
  • Consider a minimum real-word / syllable requirement for freehand text
  • Apply on the same paths as today’s validator (freehand + sign custom responses)

Related

ObservationContentValidator.isNonsense

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions