Blackbox AI CTF

Hand-authoring CTF challenges does not scale. Static challenge sets go stale; manual validation is slow.

A generator produces challenges from a category/difficulty spec, then an automated solver attempts the challenge to confirm it is solvable, calibrate difficulty, and reject infeasible output before publishing.

• 01Each generated challenge must be solvable by the in-tree solver
• 02Non-trivial difficulty floor enforced
• 03OS-agnostic packaging

• Used for internal red/blue training cycles.