Claude Code /security-review command vs cyber classifier blocks — tips and alternatives when the built-in review fails.
Anthropic ships /security-review in Claude Code. Paying users report the command runs until the classifier kills the session — ironic and expensive.
What /security-review does
It orchestrates a multi-step review of your repository — exactly the kind of session that accumulates security keywords fast.
Reducing blocks
Smaller scope per run, older Sonnet builds per error hints, fresh session per repo area.
When the command is unusable
Run reviews via uncensored API + your own checklist, or Icelake with OpenAI-compatible tools in Cursor.
Try uncensored AI free
No filters, no lecture, no training on your chats. Start in under a minute.
Built-in commands do not bypass classifiers — they often feed them.
FAQ
Will CVP unlock /security-review?
It may reduce blocks for approved orgs; solo dev approval is inconsistent per public reports.