agentleFS
andrmazandrmaz/spec-driven-architecture/.agents/skills/tools-and-context/SKILL.md

tools-and-context

Skill1 starsChanged 9 months agoMITView on GitHub

Provides Tambo with data and capabilities via custom tools, MCP servers, context helpers, and resources. Use when registering tools Tambo can call, connecting MCP servers, adding context to messages, implementing @mentions, or providing additional data sources with defineTool, mcpServers, contextHelpers, or useTamboContextAttachment.

  • Installs packages

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Add/remove helpers at runtime:

    line 121

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Tools and Context

Gives Tambo access to data and capabilities through tools, MCP servers, and context.

Quick Start

// Custom tool Tambo can call
const fetchUserTool = defineTool({
  name: "fetchUser",
  description: "Fetch user by ID",
  inputSchema: z.object({ userId: z.string() }),
  tool: async ({ userId }) => fetchUser(userId),
});

<TamboProvider tools={[fetchUserTool]}>
  <App />
</TamboProvider>;

Custom Tools

Register JavaScript functions Tambo can call:

import { defineTool, TamboProvider } from "@tambo-ai/react";
import { z } from "zod";

const fetchUserTool = defineTool({
  name: "fetchUser",
  description: "Fetch a user by ID",
  inputSchema: z.object({
    userId: z.string().describe("The user ID to fetch"),
  }),
  outputSchema: z.object({
    name: z.string(),
    email: z.string(),
  }),
  tool: async ({ userId }) => {
    const user = await fetchUser(userId);
    return user;
  },
});

<TamboProvider tools={[fetchUserTool]} components={components}>
  <App />
</TamboProvider>;

Tool Key Points

  • inputSchema: Zod object for parameters, use .describe() on fields
  • outputSchema: Zod schema for return value (optional)
  • tool: Function receives single object with input params
  • transformToContent: Enable rich content responses (images, formatted text)

MCP Servers

Connect to external MCP servers for tools, resources, prompts:

FeatureServer-sideClient-side
PerformanceFast (direct)Slower (browser proxies)
AuthOAuth + API keysBrowser session only
Local serversNoYes (localhost)
ConfigTambo dashboardReact code

Server-Side Setup

  1. Go to project dashboard
  2. Click "Add MCP Server"
  3. Enter URL and server type (StreamableHTTP or SSE)
  4. Complete OAuth if required

Client-Side Setup

npm install @modelcontextprotocol/sdk@^1.24.0 zod@^4.0.0 zod-to-json-schema@^3.25.0
import { TamboProvider } from "@tambo-ai/react";
import { MCPTransport } from "@tambo-ai/react/mcp";

<TamboProvider
  mcpServers={[
    {
      url: "http://localhost:8123/",
      serverKey: "local",
      transport: MCPTransport.HTTP,
    },
  ]}
>
  <App />
</TamboProvider>;

Context Helpers

Provide dynamic context on every message:

<TamboProvider
  contextHelpers={{
    currentPage: () => ({ url: window.location.href }),
    currentTime: () => ({ time: new Date().toISOString() }),
    selectedItems: () => selectedItems.map((i) => i.name),
  }}
>
  <App />
</TamboProvider>

Dynamic Context Helpers

Add/remove helpers at runtime:

const { addContextHelper, removeContextHelper } = useTamboContextHelpers();

useEffect(() => {
  addContextHelper("project", () => ({ projectId, projectName }));
  return () => removeContextHelper("project");
}, [projectId, projectName, addContextHelper, removeContextHelper]);

Context Attachments

One-time context for the next message (cleared after sending):

const { addContextAttachment, attachments, removeContextAttachment } =
  useTamboContextAttachment();

function handleSelectFile(file) {
  addContextAttachment({
    context: file.content,
    displayName: file.name,
    type: "file",
  });
}

Local Resources

Register @ mentionable resources users can reference in messages:

Static Resources

import { TamboProvider, ListResourceItem } from "@tambo-ai/react";

const resources: ListResourceItem[] = [
  { uri: "docs://api", name: "API Reference", mimeType: "text/plain" },
  { uri: "docs://faq", name: "FAQ", mimeType: "text/plain" },
];

const getResource = async (uri: string) => {
  const content = await fetchDoc(uri);
  return { contents: [{ uri, mimeType: "text/plain", text: content }] };
};

<TamboProvider resources={resources} getResource={getResource}>
  <App />
</TamboProvider>;

Dynamic Resources

const listResources = async (search?: string) => {
  const docs = await fetchDocs();
  return docs
    .filter((d) => !search || d.name.includes(search))
    .map((d) => ({
      uri: `docs://${d.id}`,
      name: d.title,
      mimeType: "text/plain",
    }));
};

const getResource = async (uri: string) => {
  const doc = await fetchDocument(uri);
  return { contents: [{ uri, mimeType: "text/plain", text: doc.content }] };
};

// Both listResources and getResource must be provided together
<TamboProvider listResources={listResources} getResource={getResource}>
  <App />
</TamboProvider>;

Programmatic Registration

const { registerResource, registerResources } = useTamboRegistry();

// Single resource
registerResource({
  uri: "user://file.txt",
  name: "File",
  mimeType: "text/plain",
});

// Batch registration
registerResources(
  docs.map((d) => ({
    uri: `docs://${d.id}`,
    name: d.title,
    mimeType: "text/plain",
  })),
);

Context Types Summary

TypeWhen CalledUse Case
Context HelpersEvery messageAmbient state (current page, time)
Context AttachmentsNext message onlyUser-selected files, selections
ResourcesWhen @ mentionedDocumentation, searchable data

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.