TextAI Agents
AgentAtlas Trajectory Eval Architect
Evaluate AI agents by their control decisions and trajectory quality using a six-state taxonomy and failure taxonomy, not just final outcomes.
Tag Collection
轨迹分析 Prompt Collection
3 published prompts tagged “轨迹分析”. Browse by scenario and copy in one click.
Related prompts
3 prompts
TextAI Agents
Web Agent Failure Diagnostician
Based on the three-layer framework from arXiv 2603.14248 (April 2026) — High-level Planning, Low-level Grounding, and Replanning — this diagnostician localizes failures in GUI/web agent trajectories to provide targeted, actionable fixes rather than generic improvements.
TextAI Agents
Agent Trajectory Triage Specialist
Design a lightweight, signal-based filtering system to identify high-value agent execution traces from production-scale logs for evaluation, debugging, skill mining, or safety review—without requiring ground-truth labels.
What are “轨迹分析” prompts?
They are reusable LLM prompt templates labeled with “轨迹分析” in Easy Prompt, selected for practical workflows and clear structure.
How should I use these prompts?
Open a prompt, adjust variables or constraints for your context, then copy it into ChatGPT, Claude, or your internal model.
Are all prompts on this page indexable?
This page lists published prompts with the tag. Individual bulk-synced items may still be noindex; prefer structured templates with scoring rubrics when evaluating quality.