Sandboxed code execution for agents — run Python, JavaScript, and shell scripts safely with full output capture.
Execute shell commands, manage processes, install packages, and interact with the system safely.
Long-term memory for agents — store, index, and retrieve information across sessions with semantic search.
Give your agent the ability to navigate websites, extract content, fill forms, and interact with web UIs autonomously.
Analyze datasets, compute statistics, generate visualizations, and produce data-driven insights.
Manage GitHub Actions workflows — trigger runs, monitor status, download logs, and debug failures.
Autonomous git operations — branching, committing, PR creation, code review, and merge conflict resolution.
Call any REST or GraphQL API with automatic authentication, retries, and response parsing.
End-to-end browser testing with Playwright — write, run, and debug tests across Chromium, Firefox, and WebKit.
Multi-step research skill that searches, reads, synthesizes, and produces cited reports from the open web.
AWS S3 bucket operations — upload, download, list, presign URLs, and manage objects with full IAM support.
Parse and extract structured data from PDFs, Word docs, spreadsheets, and images with OCR.
Query SQL and NoSQL databases, inspect schemas, run migrations, and generate reports from live data.
Store and search vector embeddings with Pinecone, Weaviate, or Qdrant — semantic similarity at scale.
Read, write, organize, and transform files across local and cloud storage systems.