Automate DSPy GEPA evaluation and optimization of agent skills from YAML datasets.
npx skills add https://github.com/Qredence/skills --skill dspy-gepa
Please help me install this Agent Skill. Skill: dspy-gepa Source: https://github.com/Qredence/skills/tree/main/skills/dspy-gepa Command: npx skills add https://github.com/Qredence/skills --skill dspy-gepa
Evaluates and optimizes agent skills using a DSPy-powered GEPA loop to automatically generate, test, and improve skill prompts and scenarios.
Run the GEPA workflow to generate or optimize scenarios using the provided scripts.
High-intent search queries and answers about installing and using this skill.
DSPy skill optimization uses a Generate/Evaluate/Propose/Apply loop to automatically test and refine agent prompts. It loads scenario YAML files as datasets, scores outputs with pattern-matching metrics, and iterates via teleprompter-based optimization.
The GEPA loop in DSPy is a Generate, Evaluate, Propose, and Apply cycle for agent skills. It generates scenarios, evaluates outputs against pattern-matching metrics, proposes refinements, and applies improvements to skill prompts.
DSPy teleprompter optimization requires Python 3.10 or higher. You also need the dspy-ai, PyYAML, and jsonschema dependencies installed to execute the workflow and process YAML datasets.
BootstrapFewShot and MIPROv2 teleprompters optimize prompts from YAML scenario datasets. The workflow loads these files as DSPy datasets, scores them with pattern-matching metrics, and applies the teleprompter to refine skill outputs.
Generating new scenario YAML files from skill descriptions accelerates testing and coverage. The workflow reads SKILL.md, references, and scripts to automatically generate DSPy-powered scenarios suitable for evaluation.
Evaluating DSPy agent outputs with pattern-matching metrics involves loading scenario YAML files as datasets and scoring responses against defined patterns. The GEPA loop then iterates on these scores using teleprompter-based optimization.
OmidZamaniBuild and optimize RAG pipelines with ColBERTv2 in DSPy.
TymorIbrahimCreate and manage hookify rules with YAML configuration and regex matching.
OmidZamaniOptimize Haystack pipeline prompts with DSPy and custom metrics.
jleechanorgOptimize DSPy prompts and pipelines using teleprompters and training data.
