benchmark-driven-improvement

Official

Systematic benchmark fixes for Serf.

Authorprime-radiant-inc
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Diagnoses and iterates on Serf benchmark failures to improve autonomous agent reliability.

Core Features & Use Cases

  • Provides a structured workflow for extracting, reproducing, and fixing benchmark failures.
  • Facilitates transcript analysis, session interrogation, and iterative code/prompt improvements to generalize fixes across tasks.
  • Supports local execution, tool usage assessment, and verification through standard evaluation tooling.

Quick Start

Run a local benchmark investigation workflow against a cached task, then review transcripts and iterate on prompts, tools, and code to improve robustness.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: benchmark-driven-improvement
Download link: https://github.com/prime-radiant-inc/serf/archive/main.zip#benchmark-driven-improvement

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.