Cybersecurity pro says Claude guardrails block legit work, asks about Astra
GVTZ59 · reddit · 2026-09-06
A longtime Claude user in cybersecurity says he mainly uses Claude to code custom tools for work, but Fable's restrictive guardrails make it nearly unusable, and even Opus 4.5 frequently flags his legitimate work — sometimes stopping before finishing reading his handoff memory file. His CVP requests to Anthropic keep getting rejected.
He asks how OpenAI's guardrails handle cyber-related work and whether Astra is worth trying, highlighting the real friction between model safety guardrails and legitimate security work.
More from Models
- Ornith 1.5 35B A3B beats Flint on hallucination and overthinking in user tests — BarberIcy366 · 2026-10-23
- Meta researcher: horrifying infra, not talent, is why OSS can't catch closed frontier — A_K_Nain · 2026-10-23
- TypeSafe's Jev: a "Decision Model" LLM That Returns Floats at $0.042/M Input Tokens — ariG23498 · 2026-10-23
- Microsoft's MAI-Base-2 scaling ladder pegs the L42 run at ~1.8×10^23 FLOPs — teortaxesTex · 2026-10-23
- Local model agentic coding benchmark: 27B matches 125B, tool-calling is the lever — Joralk_ · 2026-10-23
- NVIDIA open-sources PersonaPlex-7B, a full-duplex voice AI model that listens and talks simultaneously — solyarisoftware · 2026-10-23