dspy-skills — Installieren & Live-Demo

Warum nutzen

Hauptfunktionen

DSPy signatures / modules / predictors encoded as patterns
Teleprompter (BootstrapFewShot, MIPRO) usage guidance
Evaluation + metric design
RAG pipeline scaffolds with DSPy retrievers
Opinionated integration with common LLM providers

Live-Demo

In der Praxis

dspy-skill.replay ▶ bereit

0/0

Installation

Repository des Autors · Offizielle Dokumentation zu Claude Skills

Anwendungsfälle

Praxisnahe Nutzung: dspy-skills

How to build your first DSPy program and optimize it

👤 ML engineers and applied researchers ⏱ ~90 min advanced

Wann einsetzen: You have a task where prompt quality matters and want a systematic way to improve it.

Ablauf

Define the signature

Task: classify support tickets into {billing, technical, account}. Give me a DSPy signature and a simple Predict module.✓ Kopiert

→ Signature + module code
Write an eval

Add an evaluation set of 50 labeled examples and an accuracy metric.✓ Kopiert

→ Eval harness with metric callable
Optimize

Run BootstrapFewShot to compile the module against the eval set.✓ Kopiert

→ Compiled predictor + improved score

Ergebnis: A DSPy-optimized predictor that beats a hand-written prompt, with reproducible code.

Fallstricke

Eval too small — optimizer overfits — Minimum 100–200 examples; hold out a true test set
Metric doesn't capture what you care about — Spend on metric design before on model choice

Kombinieren mit: filesystem

Build a RAG pipeline with DSPy

👤 Engineers building retrieval-augmented systems ⏱ ~120 min advanced

Wann einsetzen: You want modular, optimizable RAG rather than a hand-wired chain.

Ablauf

Define the modules

Create a DSPy RAG pipeline: RetrieveThenRead with ColBERTv2 or a local retriever.✓ Kopiert

→ Modular pipeline with separate retrieval and generation
Optimize end-to-end

Write an eval on our QA set and run MIPRO to improve.✓ Kopiert

→ Compiled pipeline with score delta

Ergebnis: A RAG pipeline you can improve by changing the eval, not by rewriting prompts.

Fallstricke

Retriever quality caps end-to-end quality — Evaluate retrieval separately (recall@k) before optimizing generation

Kombinieren mit: local-rag

Kombinationen

Mit anderen MCPs für 10-fache Wirkung

dspy-skill + local-rag

Plug a local retriever into DSPy's RAG modules

Swap the ColBERT retriever for a local-rag MCP as the retrieval source.✓ Kopiert

dspy-skill + filesystem

Organize DSPy programs, evals, and artifacts in a repo

Lay out a DSPy project with programs/, evals/, and artifacts/ directories.✓ Kopiert

Werkzeuge

Was dieses MCP bereitstellt

Werkzeug	Eingaben	Wann aufrufen	Kosten
signature-design	task spec	Start of any DSPy program	0
module-authoring	signatures + flow	After signatures	0
teleprompter-optimization	module + eval	After eval is ready	LLM tokens during optimization
evaluation-harness	task data	Before optimizing	0

Kosten & Limits

Was der Betrieb kostet

API-Kontingent: LLM tokens dominate during optimization runs
Tokens pro Aufruf: Can be high — optimizations may invoke the LLM hundreds of times
Kosten in €: depends on provider
Tipp: Use cheap models during teleprompter runs, upgrade only for final evaluation

Sicherheit

Rechte, Secrets, Reichweite

Credential-Speicherung: LLM provider keys in env vars

Datenabfluss: LLM provider endpoints

Fehlerbehebung

Häufige Fehler und Lösungen

Teleprompter seems to make things worse

Check metric correctness; use a held-out set; widen the example pool.

Optimization eats your budget

Cap max_bootstrapped_demos and use cheaper models during search.

Alternativen

dspy-skills vs. andere

Alternative	Wann stattdessen	Kompromiss
prompt-architect-skill	You want prompt-level craft, not DSPy's programmatic approach	Hand-crafted vs optimized

Mehr

Ressourcen

📖 Offizielle README auf GitHub lesen

🐙 Offene Issues ansehen

🔍 Alle 400+ MCP-Server und Skills durchsuchen