● Covered by 1 source · 1 reportLow impact1 neutral
AWS Developer Advocate Proposes Enhancements for MCP Server-Agent Interactions
Key points
- MCP servers should guide agents on next steps, not just return data.
- Proposed changes include adding session and user prompt IDs for tracing.
- AWS uses these methods internally and opened its MCP Server in May.
- The goal is to improve agent task completion and tool optimization.
Rethinking MCP Server Functionality
James Ward, a principal developer advocate at AWS and a member of the Agentic AI Foundation's technical committee, advocates for a shift in how MCP servers interact with agents. Currently, many MCP servers function as simple data providers, responding to agent requests by returning data. Ward suggests that servers should act as a more active participant in the conversation, offering guidance on what an agent might do next after receiving data.
Integrating Guidance with Data Responses
This proposed change means that tool responses would include not only the requested data but also suggestions for subsequent actions and pointers to additional resources. Ward envisions MCP servers as additional voices in an agent's operational flow, providing context and direction to improve the agent's decision-making process.
Enhancing Visibility and Traceability
Another key aspect of Ward's proposal is the introduction of better visibility and tracing capabilities within MCP interactions. Unlike web development, where user actions can be traced to the database, MCP lacks a built-in trace ID. AWS addresses this by adding session ID parameters to tools, which agents pass with every call, allowing individual calls to be stitched together into a complete journey. A second parameter can carry the user's prompt or goal summary.
This tracing mechanism allows teams to compare user intent with agent execution paths, build evaluations, refine tool descriptions, and measure improvements in task completion rates. AWS has implemented much of this internally for its own MCP servers.
AWS's MCP Server and Future Directions
AWS made its MCP Server publicly available in May and added Lambda diagnostics in September. Ward previously highlighted how observing full agent journeys can reveal inefficiencies, such as an agent fetching customer records one at a time due to a server only offering a single-record tool, leading to the development of batch tools.
Ward notes that MCP is not the sole solution, as AWS also provides CLIs and skills. The optimal choice depends on the model, tool descriptions, and agent tool discovery methods, which AWS evaluates through testing.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
James Ward, a principal developer advocate at AWS, suggests that MCP servers should provide agents with guidance on subsequent actions rather than just data. This approach aims to improve agent efficiency and enable better tracing of agent interactions, moving MCP beyond a basic API to a more integrated interface.