{"version":"2.1.265","anchor":"artifact-comment-threads-get-stronger-prompt-injection-frami","canonical_anchor":"artifact-comment-threads-get-stronger-prompt-injection-frami","heading":"Artifact comment threads get stronger prompt-injection framing","tier":"internal","area":"Artifacts","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.265\/e\/artifact-comment-threads-get-stronger-prompt-injection-frami","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.265","markdown":"### Artifact comment threads get stronger prompt-injection framing\n\nArtifact comment threads now carry an explicit warning that comment text is untrusted and can't expand what Claude is allowed to do\n\n**What**\n\nThe hidden header text wrapping artifact comment threads now spells out more clearly that:\n\n- comment text is untrusted, since it's written by artifact viewers\n\n- those viewers may be outside your organization\n\n- a request embedded in a comment cannot widen the current task or grant new permissions, such as running commands, following links, touching unrelated files or configuration, or sending data or credentials\n\n**Why**\n\nThis strengthens protection against prompt injection: someone leaving a comment on an artifact can't use that comment to trick Claude into doing more than the task actually calls for, even if the comment is phrased as an instruction.\n\n- Area: Artifacts\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5"}