trace_start
Start a new trace session to create signed audit trails for AI agent safety monitoring, tracking interactions and outputs for security analysis.
Instructions
Start a new trace session for an AI agent.
Args: agent: Agent name (used in filenames). trace_dir: Directory to save trace files. model: Optional model name to attach as metadata.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| agent | No | default | |
| trace_dir | No | ./traces | |
| model | No |
Implementation Reference
- src/agent_safety_mcp/server.py:265-287 (handler)The `trace_start` tool implementation, which initializes a new `Tracer` instance with an agent name, directory, and optional metadata.
@mcp.tool() def trace_start( agent: str = "default", trace_dir: str = "./traces", model: str = "", ) -> dict: """Start a new trace session for an AI agent. Args: agent: Agent name (used in filenames). trace_dir: Directory to save trace files. model: Optional model name to attach as metadata. """ global _tracer meta = {} if model: meta["model"] = model _tracer = Tracer(agent=agent, trace_dir=trace_dir, meta=meta if meta else None) return { "status": "started", "agent": agent, "trace_dir": trace_dir, }