3. Data, Science & AI

Provides command-line interface for safely operating agentpack within Claude Code using /ap-* commands and JSON output.

Computes text statistics including word count, character count, and sentence structure metrics.

Analyzes agent-user interaction transcripts to identify context network maintenance needs and guidance improvements for AI agents.

Verifies claims in AI-generated content against reliable sources to detect and correct hallucinations, ensuring factual accuracy in outputs.

Builds AI agents, tools, and workflows with Mastra and Hono, including agent networks and TypeScript API server integration.

Manages the lifecycle of context networks to ensure AI agents have consistent, accessible context, preventing degradation from scattered documentation or missing information.

Provides quick access to Databricks documentation for data engineering, analytics, and machine learning tasks.

Provides access to Databricks Unity Catalog system tables for data lineage, audit logs, billing, compute, jobs, and query history.

Configures Databricks profiles and handles authentication for Databricks Connect, CLI, and SDK to enable seamless data platform integration.

Framework for testing Databricks skills, evaluating data processing workflows, and generating ground truth datasets via the GRP pipeline.

Generates AI/BI dashboards with strict SQL query validation via execute_sql to ensure data integrity before deployment.

Coordinates parallel exploration and implementation using specialized AI agents (scouts and implementers) for critical complex tasks.

Guides design and configuration of AI agent workflows, including node-based design, single-agent setup, and workforce orchestration using AgenticFlow.

Generates production-ready JSON workflows for fal.ai, enabling model chaining, multi-step AI generation, and image-to-video pipelines.

Generates images and videos using fal.ai's AI models with queue support for text-to-image and image-to-video requests.

Converts text to speech and transcribes speech to text using AI audio models for audio processing tasks.

Creates and configures specialized subagents within Claude Code for delegated task execution with custom prompts and tool permissions.

Provides an advanced sub-agent using Opus model's extended thinking for complex reasoning, architecture decisions, and priority judgment in AI workflows.

Designs multi-agent AI workflows with validation gates and iterative loops for autonomous development using LLM orchestration.

Guides AI agent configuration, security policies, project documentation, and team workflow setup for LLM-based systems.

Validates clinical trial data domains against CDISC standards using DataFrame structures for data integrity checks.

Designs and implements evaluation systems for AI agents, covering grader types, benchmarks, and production integration for coding, conversational, research, and computer-use agents.

Defines core principles for AI agent collaboration including decomposition, context management, and validation retrospection to optimize agent workflows.

Exports SDTM (Study Data Tabulation Model) domains to XPT or XML formats for clinical data standardization and interoperability.