Skip to content

    ← All posts

    The classifiers that watch an LLM agent

    · updated on

    A poisoned page makes an agent leak ~/.aws/credentials through curl. Where to put a classifier to stop it, how to set its threshold, why most alerts are false, what Claude Code, Bedrock and Azure cover, and what the LLM provider filters on its side.

    Spotted an error, something to add?

    Tell me: I read every message and correct the article when needed.

    Your name only appears below the article with this consent, and only if the remark is used. Your address is never published.