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

Goes well beyond the destructiveHint=true annotation: discloses that the stop is scale_to_zero, that worker disks lose data every stop while control plane/etcd/volumes survive, the four gate refusal conditions (unconfirmed non-production, production-labelled, 14-day activity, unmeasurable), twin-proposal return with standing_twin, the approval flow requiring billing.manage, and the window-conflict refusal with requested_window. This is unusually rich behavioral disclosure for a destructive tool.

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?

The scope, alternative routing, and consent requirement are front-loaded, which is good, but the body is a single very dense paragraph of ~300 words mixing gate logic, approval paths, window semantics, and consent. Each clause is substantive, yet the lack of segmentation makes it harder to scan than it needs to be.

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 destructive, non-idempotent proposal tool with no output schema, the description covers everything an agent needs: what gets destroyed and preserved, refusal paths and the required user messaging, twin-proposal behavior, where approval happens (portal_url vs. confirmation card), permission requirements, and consent handling.

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 the baseline is 3; the description earns above baseline by adding operational meaning: the default window (weekdays 19:00-07:00 plus weekends), that the expected saving is only recorded for that window, that a user-chosen stop_cron/start_cron is measured over its own hours, and that accept_node_local_data_loss must only be set true after explicit user agreement. These clarify the consequences of the cron and consent params beyond the schema text.

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?

States a specific verb and resource ('Propose stopping a cluster's worker servers outside working hours to save cost'), clarifies the artifact it creates (an off_hours_schedule decision, not a direct power-schedule write), and explicitly distinguishes itself from the sibling create_power_schedule.

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?

Names the exact conditions for preferring this tool over create_power_schedule (user wants a cluster off at night/weekends to save money, or acts on a get_cloud_savings recommendation) and adds a negative rule: when the gate refuses, do not fall back to create_power_schedule. Both when-to-use and when-not are explicit.

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