Agentic AT Protocol Feed Orchestrator
This challenge tasks you with building a sophisticated multi-agent social orchestrator. You will design a system that goes beyond simple chronological feeds to create agentic streams that autonomously categorize, summarize, and interact with decentralized social data. The core logic must utilize the LlamaIndex Workflow and Agent components to manage state and tool execution across the AT Protocol ecosystem. Your orchestrator will employ a team of specialists to handle trend detection, sentiment analysis, and automated response generation while maintaining a distinct personality profile for the user. By integrating OpenManus for autonomous tool discovery and NNI for workflow optimization, you will create a highly responsive social agent. The final interface should leverage Voiceflow for real-time voice interactions and Hostinger Website Builder for the agent dashboard. This challenge focuses on the transition from static social consumption to active, agent-driven engagement within open protocols.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a decentralized social agent using LlamaIndex and Claude Opus 4.6.6 that orchestrates content discovery and interaction on the AT Protocol.
How work is evaluated
Evaluates the agent's ability to process AT Protocol feeds and generate context-aware summaries and interactions.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Tool Call Validation
Verify agent correctly calls AT Protocol tools
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Synthesis Coherence
Clarity of summarized social trends • target: 8 • range: 0-10
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Master LlamaIndex Workflows to build event-driven multi-agent systems for real-time social streams
Implement Claude Opus 4.6.6 tool-calling capabilities to interact with complex AT Protocol schemas
Orchestrate OpenManus agents to autonomously discover and utilize new social API endpoints
Utilize NNI to tune the hyperparameters of agent decision-making loops for maximum engagement
Design a voice-first agent interface using Voiceflow that integrates with the LlamaIndex backend
Deploy a dynamic agent management dashboard using the Hostinger Website Builder API
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
Evaluates the agent's ability to process AT Protocol feeds and generate context-aware summaries and interactions.
Challenge input
JSON array of raw posts
Llama Index
Data framework for LLM
OpenManus
Open-source AI agent
Exa
Neural search API
Evaluated output
Markdown trend summary
- Verify agent correctly calls AT Protocol tools
- Clarity of summarized social trends • target: 8 • range: 0-10
- Synthesis Coherence target: 8
- 1 public reference case
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Llama Index · Required
- OpenManus · Optional
- Exa · Optional
Evaluation contract
- Tool Call Validation · Weight 1
- Synthesis Coherence · Weight 1
Metrics
- Synthesis Coherence · Target 8
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON array of raw posts
Output: Markdown trend summary
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.