3. Data, Science & AI

Computes metrics and scores for deterministic comparison of workflow runs, generating canonical artifacts for Proof of Invariance verification.

Command-line tool for semantic vector search on Agentuity cloud platform, requiring authentication for secure data retrieval operations.

Exports Google Docs and Sheets content to Markdown for analysis, summarization, or context loading in data workflows.

Ensures critical information survives context compaction in AI sessions by managing attention markers during long interactions.

Retrieves operational statistics for vector storage on the Agentuity cloud platform, requiring authentication for AI/ML data insights.

Enables CLI-based interaction with Model Context Protocol (MCP) servers to call tools, access resources, and retrieve prompts programmatically.

Automates processing of long text/code inputs by slicing, AI model execution per slice, and result aggregation via reproducible pipelines.

Executes data ingestion run using Skill Seeker, processing request data and returning structured result.

Validates AI outputs against logic, first principles, and reality to detect hallucinations, false completeness, and category errors.

Creates AI agents with built-in tools (search, calculator) leveraging LangGraph and AI SDK for automated task execution.

Automates conda environment configuration and activation for Claude Code workspaces.

Structures AI prompts into formal contracts by defining roles, scope, format, and decision rules to ensure consistent and controlled AI output.

Builds mental models for understanding AI system behavior, including context windows, probabilistic generation, and hallucination causes.

Provides structured frameworks for AI-assisted reasoning, including checklists, reasoning trees, and critique loops to maintain coherence during complex problem-solving.

Transforms ambiguous AI project requirements into structured problem statements with defined objectives, constraints, success metrics, and human-AI responsibility boundaries.

Extracts text, tables, and structured data from PDF documents for data processing and analysis.

Executes SQL queries on Agentuity cloud database with authentication for platform operations.

Optimizes AI prompts for developers via interactive clarification, reducing token usage and improving output reliability for development tasks.

Selects optimal human-AI collaboration modes: Automation, Augmentation, or Agency.

AI-powered validation of telemetry instrumentation in sandbox environments, confirming OTel spans, Kafka correlation headers, and OpenLineage events for data pipelines.

Checks current token usage and remaining tokens in an AI session, responding to queries about token count and context limits.

Creates custom Agent Skills for Claude Code to extend AI capabilities, enabling users to teach new functions, add specialized knowledge, and define reusable instructions.

Executes mandatory Exploratory Data Analysis (EDA) with data validation to ensure quality before main analysis tasks.

Automates evaluation of skill performance across Anthropic models (Sonnet, Opus, Haiku) by executing sub-agents and returning comparative scores (0-100).