VectorSpaceLab/AREX-Skill - Page 2

adalflow

Use AdalFlow to build, debug, evaluate, optimize, trace, and operate LLM task pipelines, RAG systems, agents, tools, and structured-output workflows.

agents-tools-and-streaming

Use AdalFlow Agent, Runner, ReActAgent, FunctionTool, ToolManager, streaming events, permissions, and MCP tools safely.

core-components-and-structured-io

Build service-free AdalFlow Component, DataClass, Prompt, parser, and structured-output pipelines.

evaluation-and-optimization

Guide AdalFlow evaluation metrics, datasets, parameters, gradients, prompt/few-shot/text-grad optimization, AdalComponent, Trainer, and optimize_anything workflows.

model-client-and-generator-workflows

Use AdalFlow ModelClient, Generator, and Embedder workflows for provider integration, prompt/model kwargs, output processors, caching, streaming basics, and no-credential fake-client tests.

retrieval-rag-and-data-pipelines

Document preprocessing, LocalDB persistence, retriever indexing, and RAG context assembly.

tracing-observability-and-configuration

Route AdalFlow setup, logging, tracing, MLflow, and debug-artifact workflows.

adelai-det

Routes AdelaiDet users through legacy-compatible setup, model config selection, training/evaluation, demos, text spotting, dataset preparation, and export/conversion workflows.

data-prep

Guides AdelaiDet COCO/PIC/LVIS/text dataset layout, semantic-mask generation, dataset registration, mapper expectations, and MEInst mask components.

demo-visualize

Guides AdelaiDet image/video/webcam demos, VisualizationDemo usage, confidence thresholds, output handling, and dataset visualization.

export-convert

Guides AdelaiDet checkpoint key conversion, optimizer stripping, FCOS/BlendMask weight migration, ONNX export, and optional deployment-runtime caveats.

setup-build

Guides legacy-compatible AdelaiDet installation, Detectron2/PyTorch/CUDA version selection, editable extension builds, and runtime smoke checks.

text-spotting

Guides AdelaiDet BAText/ABCNet text spotting, BezierAlign, text datasets, dictionaries, lexicons, and TextEvaluator workflows.

train-eval

Routes AdelaiDet config selection, training, evaluation, Detectron2 launch flags, checkpoints, and model-family run workflows.

adk-python

Use Google ADK Python to build agents, Workflow graphs, tools, runtime services, CLI apps, evaluations, deployments, and ADK repository changes.

agent-construction

Build and troubleshoot ADK Python Agent/LlmAgent definitions, model settings, modes, callbacks, schemas, and multi-agent delegation.

cli-configuration-deployment

Use ADK CLI commands, agent/app discovery, YAML configuration, local run/web/API server flows, eval/test commands, deployment commands, and safe CLI/config inspection.

evaluation-debugging

Design and run ADK evaluation, test, and debugging workflows for eval sets, JSON fixtures, event traces, sessions, and safe native verification.

repo-development

Modify the ADK Python repository itself with repo-specific setup, style, testing, docs, samples, schema generation, and review conventions.

runtime-services

Configure and troubleshoot ADK runtime services: Runner, App, sessions, memory, artifacts, plugins, telemetry, code executors, environments, event persistence, and service lifecycles.

tools-and-integrations

Bind and troubleshoot ADK Python tools, ToolContext, long-running and confirmation tools, auth, MCP/OpenAPI/Google API/cloud integrations, A2A helpers, and optional extras.

workflow-orchestration

Build and debug ADK 2.0 Workflow graphs, BaseNode/function nodes, graph routing, dynamic nodes, HITL, retry, checkpoint/resume, and workflow event flow.

adversarial-robustness-toolbox

Use Adversarial Robustness Toolbox (ART) for estimator wrappers, adversarial attacks, defences, poisoning/privacy/extraction, metrics, and certification workflows.

estimators-and-models

Choose and configure ART estimator wrappers for sklearn, PyTorch, TensorFlow/Keras, black-box, boosted tree, GPy, and regression models.

evaluation-and-certification

Compute ART robustness/privacy metrics, run evaluation objects, route SummaryWriter logging, and perform gradient checks and certification/verification workflows.

evasion-and-preprocessing

Use ART evasion attacks, preprocessing defences, and adversarial training recipes for image/tabular robustness workflows.

poisoning-inference-extraction

Plan ART poisoning, privacy inference, model inversion/reconstruction, extraction, and detector/mitigation workflows.

setup-and-backends

Install, import, optional dependency, backend-selection, and environment diagnostics guidance for ART users.

agent-lightning

Use this repo skill for Agent Lightning package tasks: authoring trainable agents, tracing rewards and spans, running LightningStore/Trainer loops, using agl CLI services, choosing examples, and troubleshooting optional backends.

agent-authoring

Author and debug Agent Lightning agents with rollout decorators, LitAgent classes, resource injection, return contracts, and single-rollout smoke checks.

cli-and-services

Use Agent Lightning agl CLI commands, store and Prometheus services, LLMProxy/vLLM service patterns, metrics, and safe endpoint checks.

examples-and-recipes

Choose, adapt, and troubleshoot Agent Lightning example workflows and optional dependency/backend recipes without reopening original repo examples.

runner-store-training

Operate Agent Lightning runners, LightningStore APIs, rollout status and retry behavior, custom algorithms, Trainer.fit, and Trainer.dev workflows.

tracing-and-instrumentation

Emit, inspect, adapt, and troubleshoot Agent Lightning spans, rewards, operation traces, OpenTelemetry/AgentOps tracers, and token-ID signals.

agent-starter-pack

Routes Agent Starter Pack tasks to the right workflow for creating, maintaining, and deploying generated agent projects.

deployment-ops

Guides CI/CD setup, deployment commands, observability, data ingestion, and Gemini Enterprise registration for Agent Starter Pack projects.

project-maintenance

Guides in-place enhancement, extraction, and version upgrades for projects generated by Agent Starter Pack.

project-scaffolding

Guides creation and discovery of new Agent Starter Pack projects from built-in, local, or remote templates.

agentscope

AgentScope repo skill for building agents, provider connectors, RAG and memory workflows, service deployments, and local or sandboxed workspaces.

agent-core

AgentScope agent, toolkit, permission, event, and local skill-loading workflows.