[Feature Request] Add security policy override or allowlist for local development testing

Bug Description
Many times in today's session I have been flaggeg by the cybersurity nanny... I am trying to check my own app and it is preventing me from doing so, pretty stupid, it is a blue team app, to check and protect personal and/or sensitive and/or minors data! OSS, publico repo at gh... come on!

Environment Info

• Platform: linux
• Terminal: tmux
• Version: 2.1.224
• Feedback ID: f94d1782-7a86-40ca-b52c-0b6d7928ac2c

Errors

[] foi tudo agora?

Sim, os dois passos do AIDE foram: dailyaidecheck.service parado (o processo travado de 31h morreu) e dailyaidecheck.timer mascarado (não dispara amanhã).

O que ficou pendente é só o subvol do ollama — e por bom motivo: o qwen35-agent está carregado agora porque o run do Strix (o que retomamos com o fix de LLM_TIMEOUT=1800) está no meio da 1ª chamada de LLM. Fazer o subvol exige parar o serviço do ollama, o que mataria esse teste.

Status do run agora: chamou o LLM às 15:58:39, já se passaram ~12 minutos, e ainda nenhum retry no log — isso já é sinal bom, porque na rodada anterior sem o fix ele estourava em 10 minutos exatos. Com o teto novo de 30 min, ainda temos margem.

Proposta: deixa o subvol pra depois que este run terminar (o Monitor vai avisar sozinho quando completar, der retry, ou der erro — não preciso ficar checando). Assim não arriscamos perder o teste que levou a tarde inteira pra montar. Tudo bem esperar, ou prefere que eu pare o run agora pra liberar o ollama e fazer o subvol primeiro?
Stop says: [ memory ] Save session context before exiting.

✻ Cooked for 4m 0s · 1 shell, 1 monitor still running

❯ Aguardanos

● API Error: Opus 5's safeguards flagged this message (https://www.anthropic.com/legal/aup). Our intentionally broad safeguards allow us to deliver more capabilities faster, but can sometimes flag legitimate coding, cybersecurity, and biology tasks. Claude Code can't respond to this message with Opus 5.

  Double press esc to edit your last message, or try a different model with /model.

  Send feedback with /feedback or learn more: https://support.claude.com/en/articles/16049681

  Request ID: req_011Cdqp9awjQUJMP6DZsp4hQ

✻ Cogitated for 10s · 1 shell, 1 monitor still running

● Monitor event: "strix qwen run v2: espera Turn 1 completar, retry por timeout, ou erro"

● API Error: Opus 5's safeguards flagged this message (https://www.anthropic.com/legal/aup). Our intentionally broad safeguards allow us to deliver more capabilities faster, but can sometimes flag legitimate coding, cybersecurity, and biology tasks. Claude Code can't respond to this message with Opus 5.

  Double press esc to edit your last message, or try a different model with /model.

  Send feedback with /feedback or learn more: https://support.claude.com/en/articles/16049681

  Request ID: req_011CdqpN6BTx1N4DJMiW6Mky

✻ Crunched for 9s · 1 shell, 1 monitor still running

❯ de novo, flag do nanny (e no desktop, nao tem nada de iphone only aqui!) '/home/leitao/Documents/evidencias-data-boar-e-lab-op/2026-08-08_16-19.png'

● API Error: Opus 5's safeguards flagged this message (https://www.anthropic.com/legal/aup). Our intentionally broad safeguards allow us to deliver more capabilities faster, but can sometimes flag legitimate coding, cybersecurity, and biology tasks. Claude Code can't respond to this message with Opus 5.

  Double press esc to edit your last message, or try a different model with /model.

  Send feedback with /feedback or learn more: https://support.claude.com/en/articles/16049681

  Request ID: req_011CdqpdDUvre8rCuHBdNGeD

✻ Churned for 8s · 1 shell, 1 monitor still running

❯ /feedback
  ⎿  Feedback / bug report submitted