Add the media MCP tools and live-verified slopbot path

2026-09-16 09:54:22 +02:00
parent c81ae0d0fe
commit d483a2fa47
9 changed files with 179 additions and 17 deletions
@@ -317,7 +317,16 @@ const READ_ONLY_TOOLS: Record<string, (args: any) => Promise<unknown> | unknown>
    get_action_log: args => requirePuppet().readActionLog(args?.limit ?? 100),
    get_puppet_config: () => requirePuppet().getConfig(),
    whitelist_list: () => ({ ownerId: requirePuppet().owner, whitelist: requirePuppet().whitelist }),
-    read_messages: args => paced(() => requireCore().readMessages(args?.channelId, args?.limit ?? 20))
+    read_messages: args => paced(() => requireCore().readMessages(args?.channelId, args?.limit ?? 20)),
+    list_media_scrapes: () => media.listScrapes(),
+    get_media_storage: () => media.refreshRetention(),
+    query_media: args =>
+        media.queryMedia({
+            kind: args?.kind,
+            scrapeId: args?.scrapeId,
+            limit: args?.limit,
+            offset: args?.offset
+        })
 };

 function buildServer(): McpServer {
@@ -749,6 +758,97 @@ function buildServer(): McpServer {
        async ({ userIds }) => jsonText({ whitelist: requirePuppet().removeFromWhitelist(userIds) })
    );

+    server.registerTool(
+        "scrape_media",
+        {
+            description:
+                "Start a media scrape of a channel. Uses slopbot for guild channels it is in, otherwise the client. Near the storage limit it fails with MEDIA_LIMIT unless confirm is true.",
+            inputSchema: {
+                channelId: z.string().describe("Channel ID to scrape"),
+                guildId: z.string().optional().describe("Guild ID; slopbot needs it to match"),
+                mode: z
+                    .enum(["links", "download"])
+                    .optional()
+                    .describe("links indexes URLs (default); download fetches attachment bytes"),
+                confirm: z.boolean().optional().describe("Accept the near-limit warning")
+            }
+        },
+        async ({ channelId, guildId, mode, confirm }) =>
+            jsonText(await media.startScrape({ channelId, guildId: guildId ?? null, mode, confirm }))
+    );
+
+    server.registerTool(
+        "list_media_scrapes",
+        { description: "List media scrapes with state, pages, messages, and media counts." },
+        async () => jsonText(media.listScrapes())
+    );
+
+    server.registerTool(
+        "get_media_scrape",
+        {
+            description: "Get one media scrape by ID.",
+            inputSchema: { scrapeId: z.string().describe("Scrape ID") }
+        },
+        async ({ scrapeId }) => jsonText(media.getScrape(scrapeId))
+    );
+
+    server.registerTool(
+        "stop_media_scrape",
+        {
+            description: "Stop a media scrape. It pauses after the current page.",
+            inputSchema: { scrapeId: z.string().describe("Scrape ID") }
+        },
+        async ({ scrapeId }) => {
+            media.stopScrape(scrapeId);
+            return jsonText({ stopped: scrapeId });
+        }
+    );
+
+    server.registerTool(
+        "get_media_storage",
+        {
+            description:
+                "Media storage usage against the limit, per-kind sizes, and the scrape list. Marks old scrapes when over the limit."
+        },
+        async () => jsonText(media.refreshRetention())
+    );
+
+    server.registerTool(
+        "query_media",
+        {
+            description: "Query media items across scrapes, newest scrape first.",
+            inputSchema: {
+                kind: z.enum(["image", "gif", "video", "audio"]).optional(),
+                scrapeId: z.string().optional().describe("Restrict to one scrape"),
+                limit: z.number().int().min(1).max(500).optional(),
+                offset: z.number().int().min(0).optional()
+            }
+        },
+        async ({ kind, scrapeId, limit, offset }) =>
+            jsonText(media.queryMedia({ kind, scrapeId, limit, offset }))
+    );
+
+    server.registerTool(
+        "get_media_item",
+        {
+            description: "Get one media item by scrape ID and item ID.",
+            inputSchema: {
+                scrapeId: z.string().describe("Scrape ID"),
+                itemId: z.number().int().min(1).describe("Media row ID inside the scrape")
+            }
+        },
+        async ({ scrapeId, itemId }) => jsonText(media.getMediaItem(scrapeId, itemId))
+    );
+
+    server.registerTool(
+        "delete_media_scrape",
+        {
+            description: "Delete a scrape and its files. Irreversible.",
+            inputSchema: { scrapeId: z.string().describe("Scrape ID") }
+        },
+        async ({ scrapeId }) => jsonText({ deleted: media.cleanMedia([scrapeId]) === 1 })
+    );
+
    return server;
 }