Reproducible evaluation for AI coding agents. Multi-turn scenarios against Claude Code, Codex, Copilot, Cursor, Gemini CLI, Goose, OpenCode, or any custom agent you plug in; verify behavior with rule checks, workspace diffs, multi-judge LLM consensus; pin reliability with pass^k variance across trials. Git worktrees, optional Docker sandbox.
Skill Claude CodeCodex
Identify Folio books with low on-hand stock so the operator can replenish before the title goes out of stock. Use whenever the user asks about books running low, low stock, restock list, what's almost gone, stock audit, or which titles need reordering.
Skill Claude CodeCodex
Customer support and order operations for the Folio bookstore. Use whenever the user asks about a book, an order, a refund, a return, store credit, or anything else customer-facing. All canonical data lives behind the folio MCP server - never invent prices, stock counts, customer details, or order history.
Skill Claude CodeCodex
Identify and surface every order currently in the processing state, so the operator can chase down stalled fulfilment. Use whenever the user asks about stalled orders, processing backlog, things stuck in processing, or orders that haven't shipped yet.
Skill Claude CodeCodex
Look up customer orders through the ordersdb MCP server. Use whenever the user asks about an order, a tracking number, a delivery, or a customer's purchase history.
Skill Claude CodeCodex
Operate the belt CLI to evaluate headless coding agents (Claude Code, Cursor, Codex, Gemini, and others) end to end. Use when the user asks to write or run eval scenarios, compare agents, score outputs with rules or LLM judges, register a new agent adapter, interpret reports or benchmark cards, or set up evals in CI.…