tau2-playbook

Official

Orchestrate τ²-bench evaluations with HarnessX.

AuthorDarwin-Agent
Version1.0.0
Installs0

System Documentation

What problem does it solve?

τ²-bench-specific benchmark guidance for calibrating and evaluating multi-domain dialogue systems within HarnessX, including domain-specific policies, tool usage patterns, and reward signaling.

Core Features & Use Cases

  • Domain-guided evaluation: integrates per-domain cheat-sheets, policy hints, and harness configurations to drive reliable scoring.
  • Tool gating and recovery: supports PhaseAwareToolFilter, LoopDetection, and ToolFailureGuard to improve stability across tasks.
  • Reusable harness recipes: ships with system_append_processor, guidance files, and baseline pipelines to accelerate experimentation.

Quick Start

Configure and run a τ²-bench evaluation within HarnessX to benchmark multi-domain dialogue performance with domain-specific tool usage and reward signals.

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: tau2-playbook
Download link: https://github.com/Darwin-Agent/HarnessX/archive/main.zip#tau2-playbook

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 510,000+ vetted skills library on demand.