--- title: Agent-Ready Tools Glossary: MCP, Tool Search, Pass/N, AX description: Plain definitions of MCP and agent-testing terms, grounded in real captures: deferred tools, tool search, code mode, peak context, pass/N, cost per success, AX. canonical: https://spiritroom.app/glossary/ published: 2026-10-21 updated: 2026-10-21 --- > Markdown version of https://spiritroom.app/glossary/ for agents. Index: /llms.txt [Blog](/blog/) / [Agent-ready tools](/guides/agent-ready-tools/) # Agent-ready tools glossary: tool search, deferred tools, code mode, peak context, pass/N, AX Short definitions, each with what it looked like in our own captures and runs, and a link to the page that goes deeper. Leonid Bugaev · Published 21 Oct 2026 **TL;DR** - **Loading:** deferred tools, tool search, code mode, preloading. - **Context:** peak vs cumulative, context pollution, output limits. - **Measurement:** pass/N, Wilson range, ground truth, cost per success, discovery rate. - **Design:** server instructions, annotations, two-step confirm, next-step hints, AX. [How agents load tools](#loading)[Context and cost](#context)[Measurement](#measure)[Server design](#design)[FAQ](#faq) ## How agents load tools tools/list : The MCP call that returns a server's tools with names, descriptions, input schemas and annotations. One 99-tool list we measured was 68,730 tokens. Preloading : Sending every tool definition to the model with every request. OpenCode 1.18 did this: +73,415 tokens per request for 99 tools. See [OpenCode 1 vs 2](/blog/opencode-mcp-tool-loading/). Deferred tools : Tools not sent up front. In Claude Code the model got a reminder listing 113 names only, then loaded definitions when it chose them. See [Claude Code tool search](/blog/claude-code-tool-search-explained/). Tool search : A built-in tool the model calls to find and load deferred tools: by keyword, or with `select:` and exact names. Grok Build uses BM25 keyword ranking. Code mode : Codex's default: MCP tools sit behind one code-running tool, and the model writes JavaScript to filter `ALL_TOOLS` and call them. See [how Codex finds MCP tools](/blog/how-codex-finds-mcp-tools/). alwaysLoad : A Claude Code `_meta` flag that sends one tool's definition up front; it cut tokens before the first call from 15–28k to 6–7k in our test. Discovery rate : Runs where the agent found and called your tools ÷ all runs. Codex: 34/40 in one live round. **Want these measured on your server?** We run your users' jobs through the agent apps they use and report every metric on this page, with ranges. [Test your MCP →](/#contact) ## Context and cost Peak context : The largest single model request in a run: how full the window got. Codex averaged 18.1% of its 258,400-token window on our live jobs. Cumulative input : All model requests in a run added up; roughly what's billed. About 210–218k tokens per simple question in our live runs. See [MCP context window usage](/blog/mcp-context-window-usage/). Context pollution : Tokens in tool results that didn't help: identical repeats, oversized errors, refusal boilerplate, the agent's own logged calls, unused records. See [how much MCP output is waste](/blog/mcp-context-pollution/). Output limit : The point where an app cuts or files away a tool result: 25,000 tokens in Claude Code, about 10,000 printed tokens in Codex code mode, 51,200 bytes in OpenCode. See [output limits by agent](/blog/mcp-tool-output-limits-by-agent/). Cost per success : Input tokens over all runs of a job ÷ runs that passed. See [cost per successful task](/blog/cost-per-successful-agent-task/). ## Measurement Job : A real user request in the user's words, without naming tools. Ground truth : The right answer, from fixed tool calls made outside the agent, not from agent answers. pass/N : Passing runs out of N. We never quote a single run. Wilson range : A 95% confidence interval for a pass rate that behaves well at small N: 34/40 is 71–93%. See [why pass rates need ranges](/blog/agent-pass-rates-confidence-intervals/). Silent failure : A confident wrong answer, such as “that data isn't available” when it is. Allowlist proxy : A proxy that forwards only named read tools to a production server, for safe testing. See [test against production read-only](/blog/read-only-agent-testing-production/). ## Server design Server instructions : Text an MCP server sends at connection. Apps cut it differently: 4,096 characters in Claude Code, about 2,048 in Grok Build. See [server instructions length](/blog/mcp-server-instructions-length/). Annotations : Hints on each tool: `readOnlyHint`, `destructiveHint`, `idempotentHint`, `openWorldHint`. Codex exec refused destructive-marked tools without approval. See [tool annotations](/blog/mcp-tool-annotations-destructive-hint/). Two-step confirm : A first call that changes nothing and returns the cost and a token; a second call with the token acts. See [the two-step confirm pattern](/blog/mcp-two-step-confirm-spend/). Next-step hint : A field in a result naming the next tool to call. See [next-step hints](/blog/next-step-hints-in-tool-results/). Refusal state : What an error tells the agent to do: fix arguments, retry, or ask the user. See [agent refusal states](/blog/agent-refusal-states/). Agent experience (AX) : How well AI agents do your users' jobs with your product. See [agent experience (AX)](/guides/agent-experience-ax/). **Go deeper:** [Agent-ready tools](/guides/agent-ready-tools/), [Test MCP servers with agents](/guides/test-mcp-servers-with-agents/), [How agents load MCP tools](/guides/how-agents-load-mcp-tools/). ## FAQ ### What are deferred tools? MCP tools an agent app doesn't send to the model up front. The model sees only names (or nothing) and loads definitions on demand. In our capture, Claude Code listed 113 names and loaded 3 definitions when needed. ### What is pass/N? The number of runs that passed out of N runs, such as 34/40, always shown with a confidence range, because agents vary from run to run. ### What is agent experience (AX)? How well AI agents can use your product to do your users' jobs: find it, use it right, finish, at a sane cost, safely. **Want these measured on your server?** We run your users' jobs through the agent apps they use and report every metric on this page, with ranges. [Test your MCP →](/#contact) ## Keep reading - Up[How to make your MCP server or API agent-ready, with proof (guide hub)](/guides/agent-ready-tools/) - Related[How agent apps load MCP tools: what the model sees (guide)](/guides/how-agents-load-mcp-tools/) - Related[How to test an MCP server with real AI agents (guide)](/guides/test-mcp-servers-with-agents/) - Related[Agent experience (AX): how to measure it (guide)](/guides/agent-experience-ax/)