Evaluate agent multi-turn data analysis skills using the LongDS-Bench benchmark.
npx skills add https://github.com/zjunlp/DataMind --skill longds-bench
Please help me install this Agent Skill. Skill: longds-bench Source: https://github.com/zjunlp/DataMind/tree/main/longds/runners/agent_agnostic/longds_bench Command: npx skills add https://github.com/zjunlp/DataMind --skill longds-bench
This Skill provides a platform for self-evaluating an agent's multi-turn data analysis capabilities using the LongDS-Bench benchmark.
Run the longds-bench skill to evaluate your agent's performance on the LongDS-Bench benchmark.
High-intent search queries and answers about installing and using this skill.
To evaluate multi-turn data analysis skills, you can use the LongDS-Bench benchmark to score your agent's performance. This skill provides a standardized platform for researchers to assess agent capabilities without needing Docker.
The LongDS-Bench benchmark is used for self-evaluating an agent's multi-turn data analysis capabilities. It applies to agent development workflows and provides a standardized dataset to score performance.
No, you do not need Docker to run this multi-turn data analysis benchmark. The skill allows you to run and score your agent locally without using DSGym's Docker executor, simplifying the evaluation process.
Yes, you can evaluate any custom agent harness for data analysis. The benchmark is agent-agnostic and suitable for any harness that includes a shell or code-execution tool for running the scripts.
To run the LongDS-Bench evaluation scripts, you need Python installed for script execution and a local LongDS dataset. Your agent must also have a shell or code-execution tool available.
To score an agent on standardized data analysis tasks, run the skill to evaluate your agent's performance on the LongDS-Bench benchmark. It applies the benchmark scripts to assess multi-turn capabilities.
digital-stoic-orgAutomate monorepo-aware dependency installation across Python, JavaScript, and system packages.
astoreyaiAnalyze WW memory systems and code architecture to identify weaknesses.
0xHoneyJarExecute Loa framework evaluation suites and output JSON results.
vercel-labsExecute Vercel plugin benchmark scenarios in isolated Vercel Sandboxes.
