Support an additional external agent import source

codex-rs/app-server-protocol/src/protocol/v2/config.rs

@@ -684,14 +684,19 @@ pub struct ExternalAgentConfigDetectParams {
     /// Zero or more working directories to include for repo-scoped detection.
     #[ts(optional = nullable)]
     pub cwds: Option<Vec<PathBuf>>,
+    /// Source product to detect. Missing or unrecognized values use Claude Code for backwards
+    /// compatibility; `cursor` selects Cursor.
+    #[ts(optional = nullable)]
+    pub source: Option<String>,
 }
 
 #[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema, TS)]
 #[serde(rename_all = "camelCase")]
 #[ts(export_to = "v2/")]
 pub struct ExternalAgentConfigImportParams {
     pub migration_items: Vec<ExternalAgentConfigMigrationItem>,
-    /// Source product that produced the migration items. Missing means unspecified.
+    /// Source product that produced the migration items. Use `cursor` for Cursor items; missing or
+    /// unrecognized values use Claude Code for backwards compatibility.
     #[ts(optional = nullable)]
     pub source: Option<String>,
 }

codex-rs/app-server/README.md

@@ -239,8 +239,8 @@ Example with notification opt-out:
 - `windowsSandbox/setupStart` — start Windows sandbox setup for the selected mode (`elevated` or `unelevated`); accepts an optional absolute `cwd` to target setup for a specific workspace, returns `{ started: true }` immediately, and later emits `windowsSandbox/setupCompleted`.
 - `feedback/upload` — submit a feedback report (classification + optional reason/logs, conversation_id, and optional `extraLogFiles` attachments array); returns the tracking thread id.
 - `config/read` — fetch the effective config on disk after resolving config layering, including opaque `desktop` values stored in `config.toml`.
-- `externalAgentConfig/detect` — detect migratable external-agent artifacts with `includeHome` and optional `cwds`; each detected item includes `cwd` (`null` for home), and multi-item migrations may additionally include structured `details` with plugin ids, skill names, session metadata, or other artifact names.
-- `externalAgentConfig/import` — apply selected external-agent migration items by passing explicit `migrationItems` with `cwd` (`null` for home) and any `details` returned by detect. Callers may pass `source` to identify the product that initiated the import; omitted or `null` means unspecified. The response acknowledges the synchronous import phase with an `importId`. Expected migration failures are reported as per-item failures rather than JSON-RPC errors, so the server still returns that `importId` and emits `externalAgentConfig/import/completed` with the same ID once all synchronous and background work finishes. The completion notification contains type-level `itemTypeResults` with successes and failures, including raw failure messages for the client to report separately.
+- `externalAgentConfig/detect` — detect migratable external-agent artifacts with `includeHome`, optional `cwds`, and optional `source`. Omitted or unrecognized `source` values preserve the existing Claude Code behavior; `source: "cursor"` detects portable Cursor rules, skills, commands, subagents, and MCP configuration. Each detected item includes `cwd` (`null` for home), and multi-item migrations may additionally include structured `details` with plugin ids, skill names, session metadata, or other artifact names.
+- `externalAgentConfig/import` — apply selected external-agent migration items by passing explicit `migrationItems` with `cwd` (`null` for home) and any `details` returned by detect. Pass the same `source` used for detection so the importer reads the correct product paths; omitted, `null`, or unrecognized values use Claude Code for backwards compatibility. The response acknowledges the synchronous import phase with an `importId`. Expected migration failures are reported as per-item failures rather than JSON-RPC errors, so the server still returns that `importId` and emits `externalAgentConfig/import/completed` with the same ID once all synchronous and background work finishes. The completion notification contains type-level `itemTypeResults` with successes and failures, including raw failure messages for the client to report separately.
 - `config/value/write` — write a single config key/value to the user's config.toml on disk; dotted paths such as `desktop.someKey` use the same generic write surface.
 - `config/batchWrite` — apply multiple config edits atomically to the user's config.toml on disk, with optional `reloadUserConfig: true` to hot-reload loaded threads, including multiple `desktop.*` edits.
 - `configRequirements/read` — fetch loaded requirements constraints from `requirements.toml` and/or MDM (or `null` if none are configured), including allow-lists (`allowedApprovalPolicies`, `allowedSandboxModes`, `allowedWebSearchModes`), the layered permission-profile allow map (`allowedPermissionProfiles`), the managed permission-profile default (`defaultPermissions`), lifecycle hook lockdown (`allowManagedHooksOnly`), remote-control policy (`allowRemoteControl`; `false` force-disables remote control while `true` or `null` preserves existing behavior), computer use policy (`computerUse`), pinned feature values (`featureRequirements`), managed lifecycle hooks (`hooks`), `enforceResidency`, managed new-thread defaults (`models.newThread.model`, `models.newThread.modelReasoningEffort`, and `models.newThread.serviceTier`), and `network` constraints such as canonical domain/socket permissions plus `managedAllowedDomainsOnly` and `dangerFullAccessDenylistOnly`.