runs

Browse and compare pipeline runs and archived analyses with metadata.

16|7|Updated Apr 1, 2026
One-click install
npx skills add https://github.com/ai-analyst-lab/ai-analyst-plus --skill runs-ai-analyst-lab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: runs
Source: https://github.com/ai-analyst-lab/ai-analyst-plus/tree/main/.claude/skills/runs
Command: npx skills add https://github.com/ai-analyst-lab/ai-analyst-plus --skill runs-ai-analyst-lab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Browsing and managing analytical work across multiple runs can be noisy and error-prone. This skill provides a centralized view of both active pipeline runs and archived analyses, enabling users to revisit, compare, and clean up past work with confidence.

Core Features & Use Cases

  • Browse active runs under working/runs/ with per-run metadata (pipeline_id, dataset, question, status, started_at, completed_at) and an agent progress map to track completed vs total steps.
  • Inspect archived analyses in .knowledge/analyses/ for long-term validation and benchmarking, and merge results with active runs for a holistic history.
  • Compare two runs, list the latest run, and clean up old runs using a guided, safe workflow.

Quick Start

To view your run history, say or type '/runs' in Claude Code.

Frequently Asked Questions about runs

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

FAQPage Schema
How do I browse and compare past pipeline runs?

To view your pipeline run history, simply type or say /runs in Claude Code to list active runs and archived analyses with per-run metadata like pipeline_id, dataset, status, and agent progress.

What is the best way to audit and benchmark archived analyses?

Auditing archived analyses is done by inspecting the .knowledge/analyses directory, which merges long-term validation and benchmarking data with active runs to deliver a complete, traceable history of your analytical work.

Can I clean up old pipeline runs using a safe workflow?

Yes, you can clean up old pipeline runs using a guided and safe workflow that surfaces clear results and metadata, helping you confidently remove outdated work without risking active runs or archived analyses.

How does the pipeline run history handle missing run state files or broken symlinks?

When missing run state files or broken symlinks occur, the pipeline run history gracefully handles the corrupted data by surfacing clear results and error messages, preventing noisy failures during browsing and comparison.

What metadata is tracked for active pipeline runs?

Active pipeline runs track per-run metadata including pipeline_id, dataset, question, status, started_at, and completed_at, plus an agent progress map showing completed versus total steps for each run.