mmp2055/Qwen3.5-35B-A3B-uncensored-bughunter-v8

Replace PentestGPT section with Bughunter Harness (multi-agent pipeline, v8 pinned)

PentestGPT fix: 'ollama:' prefix mandatory for local OpenAI-compat models

Add PentestGPT integration guide (setup + wrapper + operational rules)

Translate README to English

Add Q4_K_M model (with metadata patched: block_count=40 nextn=0)

Add vision projector

Add LoRA adapter v8 (Opción B: attention + shared_expert)

Add model card v8 (MoE, MTP disabled workaround)

initial commit