skills.rest
ddteeter
ddteeterCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-shepherd:workflow-analyzer

Analyze agent session transcripts and review comments to recommend workflow improvements.

Updated Feb 28, 2026
Search Tags:#workflow#ai agent#improvement#transcripts#guardrails#review-comments
One-click install
npx skills add https://github.com/ddteeter/agent-shepherd --skill agent-shepherd-workflow-analyzer
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-shepherd:workflow-analyzer
Source: https://github.com/ddteeter/agent-shepherd/tree/main/skills/agent-shepherd-workflow-analyzer
Command: npx skills add https://github.com/ddteeter/agent-shepherd --skill agent-shepherd-workflow-analyzer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Analyze agent session transcripts and review comment history to surface actionable workflow improvement recommendations for Agent Shepherd's insights loop.

Core Features & Use Cases

  • Read and inspect session transcripts (markdown-formatted) to extract patterns in tool usage, reasoning, and error handling.
  • Correlate transcripts with PR comments to surface recurring pain points and root causes.
  • Produce structured recommendations across six categories (tools, CLAUDE.md, skills, prompt engineering, agent behavior, recurring patterns) with concrete implementation steps.

Quick Start

Prompt the agent to ingest the current PR context and transcripts, then generate actionable workflow improvements.

Frequently Asked Questions about agent-shepherd:workflow-analyzer

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

FAQPage Schema
How do I analyze AI agent transcripts to improve workflow efficiency?▼

You can analyze AI agent transcripts by ingesting markdown-formatted session data to extract patterns in tool usage, reasoning, and error handling, which surfaces actionable workflow improvement recommendations.

What is the best way to identify recurring error patterns in agent sessions?▼

Correlating agent session transcripts with PR review comments surfaces recurring pain points and root causes, identifying error recovery patterns and recurring workflow issues across sessions.

How do I generate guardrails for AI agents from session history?▼

Generate guardrails by applying transcript analysis across decision points and error recovery to identify recurring patterns, producing structured recommendations with concrete implementation steps for CLAUDE.md and skills.

Can I use PR comments to find root causes of agent workflow failures?▼

Yes, you can use PR comments to find root causes by correlating review comment history with session transcripts, which surfaces recurring pain points and maps them to actionable workflow improvements.

What categories do workflow improvement recommendations cover for AI agents?▼

Workflow improvement recommendations cover six categories: tools, CLAUDE.md, skills, prompt engineering, agent behavior, and recurring patterns, with concrete implementation steps proposed when confidence is high.

Related Skills

Explore More Software Engineering Skills→
starrank-softstarrank-soft

agent-transcript

Parse and sanitize agent JSONL logs into Markdown transcripts for GitHub PRs.

Official
Advanced
scanbottscanbott

system-rules

Enforce guardrails for Rawgrowth OS governing data loading, output quality, security, and logging.

Community
Advanced
rancapolyrancapoly

pulse-engagement-listener

Orchestrate continuous employee voice capture and analysis across pulse surveys and feedback channels.

Community
Advanced
ReidondReidond

session-retrospective

Analyze Claude Code session history to identify recurring agent behavior problems.

Community
Advanced