← Voltar ao feed
low

Airlock Digital releases v7.1 with Agentic Control to block unauthorized AI agent actions at the endpoint

1 fonte(s) · 1 atualização(ões) · publicado 17/11/2026 10:39 · coletado há 5 horas

Airlock Digital announced Agentic Control with the release of Airlock Digital v7.1. It extends the vendor's Deny by Default application control to AI agents running on endpoints. The feature evaluates each agent-initiated action against customer policy before it executes. It blocks prohibited actions and returns the decision to the agent, with administrator context or an alternative ("Agentic Steering"), so the agent can adapt within policy. This is a product announcement with no vulnerability or active threat. It cites a Cloud Security Alliance report from April 2026 on unknown AI agents in organizations.

Detalhes técnicos

  • Discovers AI agents, LLMs and harnesses across the endpoint estate, building on the AI discovery in v7.0.
  • Evaluates actions in agent platforms such as Codex and Claude.
  • Centrally records sessions, commands, files, policy decisions, risk activity, token usage and estimated cost, and offers AI-generated session summaries.

The article is a vendor press release. No independent research was performed.

Linha do tempo

  1. helpnetsecurity.com ↗ · pub. 17/11/2026 10:39 · coletado 17/11 11:02 primeiro registro

    Airlock Digital releases v7.1 with Agentic Control to block unauthorized AI agent actions at the endpoint

    ler a análise desta fonte

    Airlock Digital announced Agentic Control with the release of Airlock Digital v7.1. It extends the vendor's Deny by Default application control to AI agents running on endpoints. The feature evaluates each agent-initiated action against customer policy before it executes. It blocks prohibited actions and returns the decision to the agent, with administrator context or an alternative ("Agentic Steering"), so the agent can adapt within policy. This is a product announcement with no vulnerability or active threat. It cites a Cloud Security Alliance report from April 2026 on unknown AI agents in organizations.

    Technical Details

    • Discovers AI agents, LLMs and harnesses across the endpoint estate, building on the AI discovery in v7.0.
    • Evaluates actions in agent platforms such as Codex and Claude.
    • Centrally records sessions, commands, files, policy decisions, risk activity, token usage and estimated cost, and offers AI-generated session summaries.

    The article is a vendor press release. No independent research was performed.