@growthbook/mcp 1.8.0 → 1.8.1

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@growthbook/mcp",
   "mcpName": "io.github.growthbook/growthbook-mcp",
-  "version": "1.8.0",
+  "version": "1.8.1",
   "description": "MCP Server for interacting with GrowthBook",
   "access": "public",
   "homepage": "https://github.com/growthbook/growthbook-mcp",

package/server/tools/features.js

@@ -140,7 +140,7 @@ export function registerFeatureTools({ server, baseApiUrl, apiKey, appOrigin, us
      */
     server.registerTool("get_feature_flags", {
         title: "Get Feature Flags",
-        description: "Lists feature flags in your GrowthBook organization, or fetches details for a specific flag by ID. Use to find existing flags before creating new ones, get a flag's current configuration and rules, or find flag IDs needed for create_force_rule or create_experiment. Single flag fetch (via featureFlagId) returns full config including environment rules. If flag is archived, suggest removing from codebase.",
+        description: "Lists feature flags with full details (rules, environments, values) or fetches a single flag by ID. Returns up to 100 flags per page. Use to inspect flag configuration, rules, and status. For a lightweight list of all flag IDs (no limit), use list_feature_keys instead.",
         inputSchema: z.object({
             project: featureFlagSchema.project.optional(),
             featureFlagId: featureFlagSchema.id.optional(),
@@ -188,17 +188,57 @@ export function registerFeatureTools({ server, baseApiUrl, apiKey, appOrigin, us
             ]));
         }
     });
+    /**
+     * Tool: list_feature_keys
+     */
+    server.registerTool("list_feature_keys", {
+        title: "List Feature Keys",
+        description: "Returns all feature flag IDs (keys only, no details) in your GrowthBook organization. Useful for discovering flag IDs when you need to check a large number of flags — for example, before calling get_stale_feature_flags. Optionally filter by project.",
+        inputSchema: z.object({
+            projectId: z
+                .string()
+                .optional()
+                .describe("Filter by project ID to only return flags in that project."),
+        }),
+        annotations: {
+            readOnlyHint: true,
+        },
+    }, async ({ projectId }) => {
+        try {
+            const queryParams = projectId
+                ? `?projectId=${encodeURIComponent(projectId)}`
+                : "";
+            const res = await fetchWithRateLimit(`${baseApiUrl}/api/v1/feature-keys${queryParams}`, {
+                headers: buildHeaders(apiKey),
+            });
+            await handleResNotOk(res);
+            const keys = (await res.json());
+            return {
+                content: [
+                    {
+                        type: "text",
+                        text: `**${keys.length} feature flag(s) found${projectId ? ` in project \`${projectId}\`` : ""}:**\n\n${keys.map((k) => `\`${k}\``).join(", ")}`,
+                    },
+                ],
+            };
+        }
+        catch (error) {
+            throw new Error(formatApiError(error, "fetching feature keys", [
+                "Check that your GB_API_KEY has permission to read features.",
+            ]));
+        }
+    });
     /**
      * Tool: get_stale_feature_flags
      */
     server.registerTool("get_stale_feature_flags", {
         title: "Get Stale Feature Flags",
-        description: "Given a list of feature flag IDs, checks whether each one is stale and returns cleanup guidance including replacement values and SDK search patterns. You MUST provide featureIds — gather them first from the user, from the current file context, or by using the get_feature_flags tool to list all flags and then searching the codebase for those flag IDs to determine which are present.",
+        description: "Given a list of feature flag IDs, checks whether each one is stale and returns cleanup guidance including replacement values and SDK search patterns. You MUST provide featureIds — gather them first from the user, from the current file context, or by using list_feature_keys to get all flag IDs and then searching the codebase for those IDs to determine which are present.",
         inputSchema: z.object({
             featureIds: z
                 .array(z.string())
                 .optional()
-                .describe("REQUIRED. One or more feature flag IDs to check (e.g. [\"my-feature\", \"dark-mode\"]). Gather IDs first from the user, from code context, or by using get_feature_flags to list flags and searching the codebase for those IDs."),
+                .describe("REQUIRED. One or more feature flag IDs to check (e.g. [\"my-feature\", \"dark-mode\"]). Gather IDs first from the user, from code context, or by using list_feature_keys to get all flag IDs and searching the codebase for those IDs."),
         }),
         annotations: {
             readOnlyHint: true,
@@ -216,7 +256,7 @@ export function registerFeatureTools({ server, baseApiUrl, apiKey, appOrigin, us
                                 "To gather feature flag IDs, try one of these approaches:",
                                 "1. **Ask the user** which flags they want to check",
                                 "2. **Extract from current file context** — look for flag IDs in the open file",
-                                "3. **Use the `get_feature_flags` tool** to list all flags, then search the codebase for those flag IDs to determine which are present",
+                                "3. **Use the `list_feature_keys` tool** to get all flag IDs, then search the codebase for those IDs to determine which are present",
                                 "",
                                 "Then call this tool again with the discovered flag IDs.",
                             ].join("\n"),