skills.rest
parcadei
parcadeiCOMMUNITY
Explore all skills by this author →
GitHub

braintrust-analyze

Analyze Braintrust tracing data from Claude Code sessions for patterns and insights.

3.9k|296|Updated Dec 23, 2025
Search Tags:#analysis#metrics#tracing#sessions#braintrust
One-click install
npx skills add https://github.com/parcadei/Continuous-Claude --skill braintrust-analyze
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: braintrust-analyze
Source: https://github.com/parcadei/Continuous-Claude/tree/main/.claude/skills/braintrust-analyze
Command: npx skills add https://github.com/parcadei/Continuous-Claude --skill braintrust-analyze

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Braintrust analysis enables retrospective reviews of Claude Code sessions to identify patterns, issues, and insights from tracing data.

Core Features & Use Cases

  • Session breakdown: Summarize tools, LLMS turns, and spans.
  • Patterns & Metrics: Extract insights about efficiency and bottlenecks.
  • Use Case: Post-mortem a complex session to improve future runs.

Quick Start

uv run python -m runtime.harness scripts/braintrust_analyze.py --last-session

Frequently Asked Questions about braintrust-analyze

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

FAQPage Schema
How do I analyze Claude Code sessions to identify patterns and bottlenecks?▼

Braintrust tracing analysis breaks down your Claude Code sessions into tool usage, LLM turns, and spans to surface efficiency patterns and bottlenecks. Run `uv run python -m runtime.harness scripts/braintrust_analyze.py --last-session` with BRAINTRUST_API_KEY configured to retrospectively review a session and extract actionable insights.

What data does Braintrust tracing capture from Claude Code?▼

Braintrust captures tool usage breakdowns, session replays, token trends, and span-level metrics from your Claude Code sessions. This tracing data enables post-mortem analysis to identify inefficiencies, detect loops, and inform skill creation or future session improvements.

What do I need to set up Braintrust tracing for Claude Code?▼

You need Braintrust API access, an active BRAINTRUST_API_KEY stored in ~/.claude/.env or your project environment, and plugin-enabled tracing in Claude Code. Once configured, the skill can ingest and analyze session traces without additional setup.

Can I use Braintrust analysis for debugging and productivity reviews?▼

Yes. Braintrust tracing is designed for post-mortem debugging, weekly productivity analysis, loop detection, and retrospective reviews of Claude Code sessions. Analyze patterns across multiple sessions to surface recurring issues and optimization opportunities.

What insights can I extract from analyzing session traces?▼

Session analysis reveals tool efficiency, LLM turn patterns, token consumption trends, and span-level performance. Use these insights for weekly productivity tracking, identifying repetitive loops, debugging complex workflows, and discovering opportunities to build specialized skills.

Related Skills

Explore More Data & Analytics Skills→
braintrustdatabraintrustdata
20

trace-claude-code

Trace Claude Code sessions, turns, and tool calls to Braintrust.

Official
Intermediate
humanerd-drewhumanerd-drew
2

seo-article-harvester

Monitor RSS feeds for SEO articles and generate weekly trend reports.

Official
Advanced
platonaiplatonai
1.1k

task-token-usage

Analyzes Claude Code session traces to report token usage per task.

Community
Intermediate
parcadeiparcadei
3.9k

trace-claude-code

Trace Claude Code sessions, turns, and tool calls to Braintrust.

Community
Advanced