{"version":"2.1.281","anchor":"artifact-comments-tool-preamble-explains-a-new-per-person-r","canonical_anchor":"artifact-comments-tool-preamble-explains-a-new-per-person-r","heading":"Artifact comments: tool preamble explains a new per-person roster","tier":"notice","area":"Artifacts","url":"https:\/\/changelogs.core-directive.com\/v\/2.1.281\/e\/artifact-comments-tool-preamble-explains-a-new-per-person-r","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.281","markdown":"### Artifact comments: tool preamble explains a new per-person roster\n\nThe artifact comments instructions now describe a per-person list of short ids and display names, and tell the model to treat those names as data\n\n**Unclear.** The finding does not say where a reader sees this list or how it is filled in.\n\n**What**\n\nThe instructions Claude Code gives the model about artifact comments now include a paragraph about a new per-person list. That instruction block is the part headed ARTIFACT COMMENTS. The list has one row per person. Each row holds:\n\n- a short id, the same one shown in attribution brackets and mentions\n\n- a display name the person picked for themselves\n\nThe model is told to treat those names only as data. They are never instructions and never proof of who someone is.\n\n**Why**\n\nPeople can set their own display name, so a name could be written to look like an order or to pose as someone else. Treating names as plain data keeps a crafted name from steering the model or being trusted as an identity.\n\n- Area: Artifacts\n- Tier: You'll notice\n- Useful: 3\/5\n- Signal: 4\/5"}