analysis

Analyze AI agent behavior transcripts with Docent SDK and DQL.

6|Updated Jan 31, 2026
One-click install
npx skills add https://github.com/TransluceAI/claude-code-plugins --skill analysis-transluceai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: analysis
Source: https://github.com/TransluceAI/claude-code-plugins/tree/main/plugins/docent/skills/analysis
Command: npx skills add https://github.com/TransluceAI/claude-code-plugins --skill analysis-transluceai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Docent Analysis enables structured evaluation of AI agent behavior using Docent, turning raw transcripts into actionable insights.

Core Features & Use Cases

  • Qualitative analysis of transcripts using the Docent SDK to surface behavioral patterns.
  • Query and compare agent_runs and transcripts with DQL to contextualize findings.
  • Plan-driven workflow with user approval prior to performing LLM analyses and generating results.

Quick Start

Provide a sample transcript to Docent to begin analyzing AI agent behavior.

Frequently Asked Questions about analysis

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

FAQPage Schema
How do I analyze AI agent behavior transcripts for qualitative patterns?

To analyze AI agent behavior transcripts, you can use the Docent SDK to query agent runs and surface qualitative behavioral patterns. This process transforms raw transcript data into structured, actionable insights using LLM requests.

What is the workflow for generating qualitative analyses from Docent results?

The workflow for generating qualitative analyses from Docent results requires user plan approval before executing. Once approved, the system uses LLM requests to process transcripts and generate cited behavioral insights.

Do I need to approve an analysis plan before the Docent SDK executes LLM requests?

Yes, you must approve a user plan before executing LLM analyses and generating results. This plan-driven workflow ensures you review the intended analysis approach before the Docent SDK processes the transcripts.

How does Docent cite sources when analyzing AI agent transcripts?

Docent cites sources directly in the generated outputs after analyzing AI agent transcripts. This ensures that the behavioral patterns and qualitative insights identified from agent runs are traceable to the original transcript data.

What is needed to start analyzing AI agent behavior using Docent?

To start analyzing AI agent behavior, you need to provide a sample transcript to Docent. The system then uses the LLMRequest and DQL components from the Docent SDK to query the data and identify behavioral patterns.