Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations give only readOnlyHint=false, openWorldHint=true, no idempotency. The description adds substantial behavioral context beyond these: invoke runs one-off without installing or bloating the tool list, credential/wallet states return connect/checkout links with an explicit retry instruction, writes require owner/admin, and publish_prompt returns a login-free share link. No contradiction with annotations. Minor gap: does not describe the output format of invoke/search results.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Every sentence carries real information and the content is front-loaded with the core purpose. However, it is a single ~300-word unbroken paragraph with no bullets, headers, or action breakdown — the 14 actions and their nuances blur together. Length is justified by complexity, but the lack of structure significantly hurts scannability for an agent parsing this description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a tool with 14 actions, 23 params, minimal annotations, and no output schema, the description is remarkably complete: the core flow, invoke edge cases (non-installed execution, connect/checkout retry loop), install-vs-invoke decision, permission requirements, installed_in_toolkit vs installed_in_workspace flags, and the entire prompt library subsystem are all covered. Gaps: exact shape of describe's profile output and mechanics of conversation/vars params. Still, it's about as complete as a hub-tool description gets without becoming bloated.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 0%, so the description carries the full burden. It meaningfully explains the core workflow parameters (action enum, mcp_id, tool_id, arguments, query, prompt slugs/vars) by mapping them into the search→describe→invoke flow and the prompt library. However, several params remain unexplained (conversation, tier_slug, immediate, request_name, cancel_reason/cancel_comment, prompt_targets). Partial compensation for the 0% coverage, but a 23-param surface deserves more param-level detail.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states this is the mcp.ai marketplace — the catalog of every MCP/tool AND the mechanism to run them. It explicitly enumerates each action's purpose (search discovers, describe returns a profile, invoke runs, install makes permanent, etc.) and distinguishes itself from sibling tools like authenticate/connect/toolkit_info. For a composite tool with 14 actions, the purpose is exceptionally well defined.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Extremely explicit usage guidance: 'prefer invoke for a single/occasional use' vs 'use install only to make an MCP PERMANENT', the auto-retry flow after connect/checkout links, and 'request_mcp asks us to build a NEW MCP when nothing fits'. It also documents the write-permission requirement (owner/admin). This is model usage guidance — when, when-not, and named alternatives.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.