analyze-trajectory

Dispatch sub-agents to analyze CI logs and diagnose recurring failures.

1.9k|124|Updated Mar 1, 2026
One-click install
npx skills add https://github.com/yologdev/yoyo-evolve --skill analyze-trajectory
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: analyze-trajectory
Source: https://github.com/yologdev/yoyo-evolve/tree/main/skills/analyze-trajectory
Command: npx skills add https://github.com/yologdev/yoyo-evolve --skill analyze-trajectory

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Diagnose recurring CI failures by dispatching sub-agents to digest CI logs and produce a concise root-cause diagnosis.

Core Features & Use Cases

  • Distributed log analysis: Offload large log reading to sub-agents to prevent bloating the main context while preserving fidelity.
  • Incremental diagnosis: Return a concise 1-3 sentence diagnosis with supporting evidence (e.g., run IDs, commit SHAs) and a path to resolution.
  • Artifact references: Document artifacts like CI run IDs and PRs to anchor the diagnosis and guide next steps.

Quick Start

Trigger analyze-trajectory when CI logs show a recurring failure to obtain a focused root-cause diagnosis.

Frequently Asked Questions about analyze-trajectory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I diagnose recurring CI failure trajectories from large logs?

To diagnose recurring CI failure trajectories, you can dispatch sub-agents to digest large CI logs and return a concise root-cause diagnosis. This approach preserves context fidelity while preventing main context bloat.

What is the best way to find the root cause of persistent CI errors and frequent reverts?

Finding the root cause of persistent CI errors involves analyzing trajectory failures with sub-agents that produce a 1-3 sentence diagnosis. This diagnosis includes supporting evidence like run IDs and commit SHAs.

Can I chunk large CI artifacts to prevent context window issues during log analysis?

Yes, you can chunk large CI artifacts during log analysis to prevent context window issues. The process uses shared_state storage for artifacts and dispatches sub-agents to handle the chunked data.

How do I ensure CI log diagnosis outputs are validated and reference concrete runs?

To ensure CI log diagnosis outputs are validated and reference concrete runs, the analysis enforces strict JSON outputs and validation rules. This guarantees evidence references point to specific run IDs or commit SHAs.

When should I use sub-agents to analyze CI logs instead of reading them directly?

You should use sub-agents to analyze CI logs instead of reading them directly when dealing with STUCK tasks or recurring failures. This offloads large log reading to prevent bloating the main context while preserving fidelity.