3. Data, Science & AI

Optimizes AgentDB performance through quantization, HNSW indexing, caching, and batch operations for memory efficiency and faster vector search at scale.

Generates coding conventions and rules for AI coding agents to ensure consistent and efficient development workflows.

Automatically optimizes AGENTS.md and agent rules for token efficiency, enhancing LLM agent performance and reducing input costs.

Provides schema design, migration, and query assistance using Drizzle ORM with PostgreSQL for API routes and backend data operations.

Crafts OpenCode rules to customize AI agent behavior by analyzing project conventions and drafting formatted rules with globs/keywords.

Creates or modifies evaluations for MCP Server, enabling AI model performance assessment and validation in machine learning pipelines.

Enables real-time web search using Google Search integrated with Gemini AI for enhanced information retrieval.

Creates customizable generic prompts for any task, enabling structured input for LLM-based applications and workflows.

Explores codebases using parallel AI agents to efficiently navigate and analyze complex code structures.

Delegates complex problems to Gemini Deep Think for advanced AI reasoning and solution generation.

Analyzes codebases and documentation using Google's Gemini AI model with 1 million token context for comprehensive insights.

Intelligently selects optimal chart type from 26 options and generates chart image using JavaScript based on user data specifications.

Generates structured narrative reports from data using T8 schema with entity labeling and semantic markup for data interpretation.

Adds new immutable constraints to Claude's constitution, enforcing consistent rule adherence across all AI responses.

Assists in selecting statistical tests, interpreting effect sizes, and conducting power analysis for data-driven research and decision-making.

Manages long-running computational tasks such as model training and large-scale data processing to prevent timeout issues.

Processes oversized files (>100KB, >2000 lines) using Python REPL analysis, LLM semantic reasoning, and subagents for synthesis.

Automates phylogenetic biogeographic analyses with BioGeoBEARS in R, handling data validation, reformatting, and visualization for species distribution studies.

Generates phylogenetic trees from genome assemblies using BUSCO and compleasm-based single-copy orthologs with scheduler-aware workflow automation.

Extracts structured data from scientific PDFs to enable systematic reviews, meta-analyses, and database creation with validation metrics.

Automates parallel investigation and resolution of independent software failures by dispatching dedicated subagents per domain.

Generates AI-powered videos using Google Veo models, supporting text-to-video, image-to-video, lip-sync, and scene extensions via CLI.

Optimizes CLAUDE.md files for Claude 4 model performance using Anthropic's best practices, enhancing prompt effectiveness and system instructions.

Provides a secure, restricted environment for executing dbt data transformation workflows on Google BigQuery data warehouse.