Tools
Browse everything AgentGraph has indexed (45,320 active tools).
MCP server for Atlassian Bitbucket Data Center - interact with repositories and code
Version control & code hostingMCP server for Atlassian Confluence Data Center - access and manage content
General capabilityMCP server for Atlassian Jira Data Center - search, view, and create issues
Productivity & project managementAgent-first prediction markets on Base. Trade, quote, create markets via MCP.
General capabilitySolana token safety for AI agents — rug-pull, honeypot & Token-2022 trap detection before you buy.
General capabilitySearch Console for agents: index status, unindexed-URL discovery, Request Indexing, sitemaps.
General capability- oraclemcpMCP
Engine-free, governed Oracle Database MCP server with a fail-closed SQL guard (unofficial).
Database querying & management Relays messages between two Claude Code sessions on different machines using the Channels API
General capabilityStructured-output enforcer: extract and validate JSON from messy LLM text.
AI / ML model accessToken-aware message truncation: fit a chat history into your model's context budget.
General capabilityNetwork-egress firewall for agent tools: check URLs against a declarative policy.
General capabilitySnapshot tests for AI tool-call traces: diff, validate, and catch silent regressions.
General capabilityValidate tool-call args before execution. Returns LLM-friendly retry hints.
AI / ML model accessInject, parse, and strip [N] citation markers in RAG outputs.
General capabilityReliable CSV parsing + generation. Handles quotes, embedded commas, BOMs, CRLF.
General capabilityCharacter-precise unified diff + patch application + patch parsing.
General capabilityConvert HTML to Markdown or strip to plain text. For web-scraping agents.
Browser automation & web scrapingRun JMESPath queries against JSON. Pure JS, no native binary needed.
General capabilityParse JSON5 (comments, trailing commas, unquoted keys) and emit strict JSON.
General capabilityToken-budget-aware text handling: count, truncate, and chunk for LLM prompts.
AI / ML model access