judge-trajectory

name	judge-trajectory
description	Evaluate agent trajectories, tool-use traces, intermediate artifacts, runtime failures, and side effects when process quality is part of the verdict. Use this whenever the judging task involves agent runs, tool calls, planning quality, web or code traces, process reliability, or any comparison where the final answer alone is not enough, even if the user only says judge the run, compare traces, or evaluate the agent workflow.
license	MIT
compatibility	Works in agents that support the Agent Skills standard. The `judge` agent should load this skill through the agent-skills MCP server when process evidence matters.
metadata	{"author":"your-org","version":"0.1.0"}

Trajectory Judging

Use this skill when the decisive evidence lives in the process: plans, tool calls, intermediate artifacts, failure handling, gathered evidence, or side effects.

Read references/trajectory-techniques.md when you need the benchmark rationale for process-aware judging, trajectory rubrics, verifier-backed evidence gathering, and skepticism toward narrated reasoning.

When to use this skill

• Agent-run evaluation where trajectories, tool traces, or intermediate artifacts matter.
• Web, code, RAG, or tool-use comparisons where process reliability is part of the verdict.