build-si-agent

Implement AdCP sponsored intelligence agents with session lifecycle and message handling.

18|15|Updated Aug 20, 2025
One-click install
npx skills add https://github.com/adcontextprotocol/adcp-client --skill build-si-agent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: build-si-agent
Source: https://github.com/adcontextprotocol/adcp-client/tree/main/skills/build-si-agent
Command: npx skills add https://github.com/adcontextprotocol/adcp-client --skill build-si-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Enables teams to implement a Sponsored Intelligence (SI) agent that manages offering discovery, session lifecycle, message exchange, and termination so platforms can serve conversational sponsored content reliably across synchronous and asynchronous AdCP flows.

Core Features & Use Cases

  • Offering discovery and tokens: check availability and return offering_token and TTL for buyers to initiate sessions.
  • Session lifecycle management: create sessions with session_id and session_status, persist state, and terminate sessions with explicit termination confirmation.
  • Message handling with sponsored responses: process user messages and return structured sponsored content responses (content, content_type) while echoing session identifiers.
  • Context passthrough and persistence: echo opaque context objects unchanged in every response and use ctx.store for durable session state.
  • Developer workflow and testing: auto-generated capabilities, storyboards for validation, and SDK patterns for building and serving SI agents.

Quick Start

Start an SI agent that issues offerings, initiates sessions, responds to messages with sponsored content, and terminates sessions while persisting state and echoing context using the AdCP SDK.

Frequently Asked Questions about build-si-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a sponsored intelligence agent for conversational sponsored content?

To build a sponsored intelligence agent, implement AdCP SI handlers for offering discovery, session lifecycle, message exchange, and termination. This Skill provides the handlers needed to serve conversational sponsored content reliably across synchronous and asynchronous flows.

How does session lifecycle management work for AdCP sponsored intelligence?

Session lifecycle management creates sessions with a session_id and session_status, persists state via ctx.store, and terminates sessions with explicit confirmation. This ensures durable state tracking and clean closure for sponsored content delivery.

Can I handle both synchronous and asynchronous message flows in sponsored intelligence sessions?

Yes, this Skill supports both synchronous and asynchronous AdCP flows. It processes user messages and returns structured sponsored content responses with content and content_type while echoing session identifiers across both communication patterns.

What is offering discovery and how do offering tokens work in AdCP?

Offering discovery checks availability and returns an offering_token with a TTL for buyers to initiate sessions. This mechanism allows platforms to manage and validate sponsored content opportunities before session creation.

Do I need to manually manage session state when building an AdCP SI agent?

No, you do not need to manually manage session state. The Skill uses ctx.store for durable session state persistence and echoes opaque context objects unchanged in every response, handling state automatically.

What is the best way to validate and test a sponsored intelligence agent implementation?

The best way to validate an SI agent is using the auto-generated capabilities and storyboards provided. These SDK patterns allow you to test offering discovery, session initiation, message handling, and termination before deployment.