Skill Claude CodeCodex
Analyze a codebase and recommend Claude Code automations (hooks, subagents, skills, plugins, MCP servers). Use when user asks for automation recommendations, wants to optimize their Claude Code setup, mentions improving Claude Code workflows, asks how to first set up Claude Code for a project, or wants to know what…
Skill Claude CodeCodex
Run a Petri-style parallel audit of an Elliot connector — spawn 5 sub-agents that try to use the connector for real, capture every failure, score the run on graded dimensions, and drive a fix loop until the connector passes. Use after a connector is built and linted, before deploying it.
Skill Claude CodeCodex
Build an Elliot connector that turns an API or database into agent-ready MCP tools. Use when the user wants to connect a new data source, wrap an existing API, or make their database queryable by AI agents.
Skill Claude CodeCodex
Compose a multi-step Elliot skill — a named sequence of connector tool calls that the runtime exposes as one MCP tool. Use when a single user-level job-to-be-done needs more than one tool to answer, and you want the agent to invoke it as a single call instead of orchestrating the chain itself.
Skill Claude CodeCodex
Deploy an Elliot connector to the runtime so agents can use it. Runs the full validation pipeline (lint → eval → deploy) before activating. Use when the user is ready to make their connector live.
Skill Claude CodeCodex
Discover a data source (REST API, Postgres, MySQL, or local file) and register it with Elliot so its tables and endpoints become queryable. Use as the first step of building a new connector.
Skill Claude CodeCodex
Master onboarding for Elliot. Read this first when you connect to Elliot — and ALSO read this when any Elliot tool call fails with a connection error, because the most likely cause is the Elliot server isn't running yet. Explains how to start Elliot, the five principles, the canonical workflow, and which Elliot…
Skill Claude CodeCodex
Lint an Elliot connector file and fix all reported issues. Use when the user wants to validate a connector, improve tool descriptions, or fix quality warnings.
Skill Claude CodeCodex
Wire the Elliot trace hook into a coding agent (Claude Code, Codex, Cursor) so its prompts, model reasoning, tool calls, and final answers stream live into the Elliot Agent Console alongside the MCP tool calls.
Skill Claude CodeCodex
Guided onboarding for turning a product into an agentic product with Elliot. Use this as the FIRST step when a user wants to make their API or database agent-ready — it interviews the user about their product and how agents should use it before any tool is designed, so the connector reflects intent instead of…
Skill Claude CodeCodex
Run Elliot eval suites to validate connector tool quality. Use when the user wants to test their connector against expected outputs, check token budgets, or verify tool correctness.