{"version":"2.1.273","anchor":"new-guidance-text-on-learning-connector-tool-schemas-before","canonical_anchor":"new-guidance-text-on-learning-connector-tool-schemas-before","heading":"New guidance for learning connector-tool schemas before publishing an Artifact or Doc","tier":"internal","area":"Artifacts","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.273\/e\/new-guidance-text-on-learning-connector-tool-schemas-before","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.273","markdown":"### New guidance for learning connector-tool schemas before publishing an Artifact or Doc\n\nClaude now has explicit guidance on how to safely learn a connector tool's argument and result shapes before publishing a page that uses it\n\n**What**\n\nA new block of guidance text instructs Claude on how to handle connector (MCP) tools when publishing something like a Claude Doc or Artifact page:\n\n- Take argument names for a loaded connector tool from the tool's own input schema, not from guesswork.\n\n- Only learn a tool's result shape by making a safe, real call to it, never by inferring it from a write call.\n\n- Use `describeTool(server, tool)` when the published page needs to read a tool's schema at view time, since the authoring session cannot see that information before publishing.\n\n- Disclose to the user, rather than embedding in the page itself, whenever a tool's shape had to be guessed.\n\n**Why**\n\nThis reduces the chance of a published page calling a connector tool with wrong argument names or misinterpreting its results, since the model is directed to verify tool shapes safely instead of guessing.\n\n- Area: Artifacts\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5"}