AgentsAvançado

W&B Primary Skill

porwandb·wandb· v1.0.0 · atualizado em 2026-04-10

Score

Comprehensive primary skill for agents working with Weights & Biases. Covers both the W&B SDK (training runs, metrics, artifacts, sweeps) and the Weave SDK (GenAI traces, evaluations, scorers). Includes helper libraries, gotcha tables, and data analysis patterns. Use this skill whenever the user asks about W&B runs, Weave traces, evaluations, training metrics, loss curves, model comparisons, or any Weights & Biases data — even if they don't say "W&B" explicitly.

wandbweights-and-biasesweavegenai-tracesevaluationtraining-metricsllm-observability

Linguagens

TypeScriptJavaScriptPythonJava

0Stars

0Forks

0Usos

Fork

Documento do Skill

SKILL.mdwandb-primary/workflow

Identify the data source: — Determine if the user is asking about training runs (W&B SDK) or GenAI traces/evaluations (Weave SDK).

Query the data: — Use the appropriate SDK (wandb or weave) to retrieve the relevant data based on the user's request.

Data Transformation: — Convert Weave wrapper types to plain Python using `weave_helpers.unwrap()` if necessary.

Load into DataFrame: — Convert the data into a pandas DataFrame for easier analysis using `wandb_helpers.runs_to_dataframe()` or similar functions.

Analyze the data: — Compute relevant statistics (e.g., mean, min, max, token usage) and identify key trends or anomalies.

Summarize findings: — Create a concise summary of the analysis, including key metrics and insights.

Present results: — Present the findings in a structured format (e.g., tables, bullet points) and answer the user's questions directly.

Telemetria de Agentes

Execuções

total

Taxa de Sucesso

últimos 30d

Latência Média

0.0s

p50

Alucinação

0.0%

detecção

Tokens Entrada

avg 0/exec

Tokens Saída

avg 0/exec

Uso por Plataforma

Skills Relacionados

Similar aDynamic Stop Loss (DSL) v4

60%

Similar aAdversarial Review

60%

Árvore do Skill

W&B Primary Skill

wandb-primary

Fases Cognitivas6

1.SENSE

2.CONTEXTUALIZE

3.HYPOTHESIZE

4.EVALUATE

5.RECOMMEND

6.REFLECT

Triggers10

analyze W&B runsquery Weave tracesevaluate model performance in W&Bcompare training metricsinspect loss curvessummarize W&B dataextract evaluation resultsget token usage from Weave tracesdiagnose a W&B runanalyze GenAI traces

Avaliar este Skill

Score Breakdown

⭐Avaliação Humana0%

🤖Sucesso de Agentes0%

🕐Atualidade100%

🔗Saúde de Dependências100%

🕸️Centralidade no Grafo0%

🛡️Segurança50%

CompositeScore = α·Humano + β·Agente + γ·Recência + δ·Deps + ε·Centralidade + ζ·Segurança

Instalação

$ synaptic mcp download wandb-primary

$ synaptic skills detail wandb-primary

$ synaptic skills live wandb-primary

Dependências

wandb weave pandas numpy

Links

GitHub Repository