Add DuoNeural community links + team credits

Q8_0 proper GGUF replaces stub

Q4_K_M proper GGUF replaces stub

Fix model card: add YAML frontmatter, prominent GGUF download section

Update README: correct GGUF sizes + vocab size note

Add ghostshell-4b-Q8_0.gguf

Add ghostshell-4b-Q4_K_M.gguf

Add GhostShell-4B BF16 merged weights — Gemma4-E4B abliterated + SFT

Add model card

initial commit