Golden Flow: FLOW-03

Single Agent – With Tools

Tool integration via agent_role field. Validates protocol semantics for tool-enabled workflows.

Goal

Validate that the runtime can handle tool-based steps with proper agent_role semantics.

Success Criteria

Tool steps execute via agent_role, outputs captured in trace.

Sequence

Execution Steps

The normative sequence of operations for this flow.

1

Tool Context

Context for API testing workflow.

2

Tool Steps

Steps with agent_role: curl_executor, jq_processor.

3

Role-Based Execution

Each role handles specific tool type.

4

Result Capture

Tool outputs recorded in trace.

Normative

Normative Scope

This flow normatively aggregates and enforces existing constraints from the following MPLP modules.

Plan Module

Steps with tool roles MUST use agent_role field (not fictional tool_name).

Extension Module

Tool invocations MUST go through protocol-defined extension points.

Trace Module

Tool execution results MUST be captured in trace events.

Validation

Failure Conditions

Execution of this flow MUST be considered non-conformant if any of the following conditions occur.

Tool invocations bypass the agent_role mechanism.

Tool outputs are not captured in trace.

Security sandbox violations during tool execution.

Non-deterministic tool behavior in golden tests.

Protocol Conformance Requirement

Any MPLP-conformant runtime is expected to pass this flow without violating any of the normative constraints defined above.

Failure to satisfy this flow results in a non-conformant MPLP implementation.

Key Modules

ContextPlanTraceExtension

Positioning Notice: This website provides discovery and positioning content only. For formal protocol definitions, see docs.mplp.io. Source of truth: GitHub Repository.