Braintrust Analyze

Name: Braintrust Analyze
Author: parcadei

🌐Community

by parcadei · vlatest · Repository

Analyzes complex situations, identifies key stakeholders, and suggests strategic actions based on diverse data inputs.

Install on your platform

We auto-selected Claude Code based on this skill’s supported platforms.

Run in terminal (recommended)

terminal

claude mcp add braintrust-analyze npx -- -y @trustedskills/braintrust-analyze

Or manually add to ~/.claude/settings.json

~/.claude/settings.json

{
  "mcpServers": {
    "braintrust-analyze": {
      "command": "npx",
      "args": [
        "-y",
        "@trustedskills/braintrust-analyze"
      ]
    }
  }
}

Requires Claude Code (claude CLI). Run claude --version to verify your install.

About This Skill

What it does

The Braintrust Analyze skill analyzes Claude Code sessions to provide insights into your workflow. It examines tracing data to identify patterns, issues, and opportunities for improvement within your AI agent interactions. Specifically, it breaks down tool usage, agent activity (like plan-agent or debug-agent), skill activations, and token consumption.

When to use it

This skill is useful in the following scenarios:

After completing a complex task to perform a retrospective analysis.
When debugging why a task failed.
For weekly reviews of your productivity patterns.
To identify opportunities for creating new skills or improving existing ones.
To understand token usage trends over time.

Key capabilities

Session Analysis: Provides detailed breakdowns of individual Claude Code sessions, including tool usage and agent activity.
Loop Detection: Identifies repeated sequences of the same tool calls, potentially indicating inefficient approaches or stuck loops.
Usage Statistics: Tracks agent and skill usage over time (last 7 days).
Token Trends: Visualizes token consumption trends.
Session Replay: Allows you to replay a specific session to view the full sequence of actions taken.

Example prompts

Here are some example commands that can be used with this skill:

uv run python -m runtime.harness scripts/braintrust_analyze.py --last-session: Analyze the last completed session and display a summary.
uv run python -m runtime.harness scripts/braintrust_analyze.py --detect-loops: Detect any looping patterns in recent sessions.
uv run python -m runtime.harness scripts/braintrust_analyze.py --weekly-summary: Generate a weekly summary of your activity.

Tips & gotchas

Prerequisites: Requires a BRAINTRUST_API_KEY configured in either your user's .claude/.env file or the project's .env file. Braintrust tracing must also be enabled through the braintrust-claude-plugin.
The skill is designed for analyzing Claude Code sessions, not general AI agent interactions.

View Repository →

TrustedSkills Verification

Unlike other registries that point to live repositories, TrustedSkills pins every skill to a verified commit hash. This protects you from malicious updates — what you install today is exactly what was reviewed and verified.

Security Audits

Gen Agent Trust Hub	Pass
Socket	Pass
Snyk	Pass

Details

Version: vlatest
License
Author: parcadei
Installs: 130

Repository (canonical source) →

🌐 Community

Passed automated security scans.