harbor - Harbor Framework for Agent Evaluation
Runs harbor commands and manages agent evaluation tasks
Tags
Updated: 2026-02-24Capabilities
Typical Inputs
Typical Outputs
What this skill does
- run harbor commands
- validate task structure
- execute agent tasks
- list datasets
- check task results
- manage local workspace
Inputs
- task files
- API keys
- dataset name
- agent name
- model identifier
Outputs
- execution log
- reward file
- test details
Requirements
- harbor installation
- task directory structure
- API key for cloud execution