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

Beyond the annotations (destructiveHint=true, readOnlyHint=false), it discloses that get_mode is read-only, that approve/reject are refused for every AI in hosted deployments, that explicit in-chat user consent must be quoted in user_request, and that deletes require the exact project name. It also warns against self-approval and mode changes made to persist the agent's own writes — real behavioral context an agent could not infer from the flags.

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?

Purpose and the safety-critical constraint are front-loaded, and nearly every clause carries information with no filler. The telegraphic fragment style ("list (a line each; item_id: one in full)") is dense to the point of being slower to parse than it needs to be, costing it the top score.

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 20-parameter, six-action tool with no output schema, the description covers the return shapes ("Returns {items, total}, the mode, or the result") and the consent/confirmation requirements. Remaining gaps are minor paging and filtering behavior, which the 100%-covered schema already documents.

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 100%, so baseline would be 3; the description earns an extra point by explaining action-gated semantics: item_id means "one in full" for list but the target id for approve/reject, keep=a|b|both resolves conflicts, and project is required to approve or reject. It does not touch a few parameters (after, limit, status, model), leaving those entirely to the schema.

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 names the resource ("the user's review queue and mode") and enumerates the concrete operations (list, walk, get_mode, approve, reject, set_mode), so an agent can tell this is the human-in-the-loop approval surface rather than a memory/history write tool. It stops short of explicitly contrasting itself with siblings like manage_history or manage_links, so it lands at 4 rather than 5.

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?

It gives per-action triggers ("walk: when the user asks what's waiting (boxes or links)"), ordering constraints ("approve: make it a record (edit / move_to first)"), and an explicit prohibition plus alternative: "ONLY call this ... when the user has explicitly told you, in this chat" and "Hosted: approve/reject are refused for every AI; share the review page link." That is genuine when/when-not/alternative guidance.

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