{"version":"2.1.299","anchor":"permission-prompt-tool-host-display-permissions-handled-d","canonical_anchor":"permission-prompt-tool-host-display-permissions-handled-d","heading":"Permission requests worded by a host app are handled differently","tier":"notice","area":"SDK","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.299\/e\/permission-prompt-tool-host-display-permissions-handled-d","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.299","markdown":"### Permission requests worded by a host app are handled differently\n\nWith `--permission-prompt-tool`, Claude Code now checks whether a permission request carries wording from the host app instead of a local-only mark\n\n**Unclear.** What this changes in which requests the prompt tool or a hook ends up answering is not clear.\n\n**What**\n\n`--permission-prompt-tool` names an MCP tool that answers permission prompts in non-interactive mode. Some applications that embed Claude Code supply their own wording for these prompts.\n\nWhen deciding how to handle a request, Claude Code now checks whether it carries wording from the host application, instead of whether it is marked for local display only. It also tells the prompt tool whether the wording came from the host, and passes the permission result on to the hook check that runs at the same time. Separately, \"ask\" decisions (where Claude Code should ask before acting) that carry host wording are now kept as asks.\n\n**Why**\n\nThis affects applications that supply their own wording for permission prompts.\n\n- Area: SDK\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 2\/5\n- Scope: individual\n- Heads-up: no"}