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

Adds real context beyond readOnlyHint/idempotentHint: nothing is mutated, a plan_id is issued, weakening changes are rejected with a dashboard pointer, and the caller is admin/owner-gated. One gap remains: the schema enumerates weakening kinds (delete_database, delete_notification_channel, manage_api_key revoke, untag_sensitive_column) without clarifying whether the tool rejects them up front or accepts-then-refuses, which matters for how the agent handles the error.

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?

Front-loaded with 'Step 1 of 2' and the read-only guarantee, then the whitelist, then the refusal rule, then auth, then the apply handoff. Dense but each clause carries information; the long parenthetical enumeration of allowed operations is the only mildly bloated part.

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 single `changes` array with ten oneOf variants and no output schema, the description covers return values (diff + plan_id), atomicity by implication via the schema note, auth level, and the follow-up tool. It falls short only on explaining how weakening kinds in the schema are actually handled.

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% with rich per-variant field descriptions, so the baseline is 3. The description does add whitelist semantics by naming which change kinds are permissible, but it does not reconcile that whitelist with the delete/revoke variants present in the schema, leaving the parameter contract partially ambiguous.

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 ('Plans one or more changes... returns their diff and a plan_id'), names it as 'Step 1 of 2', and explicitly distinguishes it from apply_config_change, the sibling that performs the write. An agent can route between the two without opening either schema.

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?

Explicit when-to-use (additive hardening: add database, enable rule, raise severity, enforce, fail closed, sanitized telemetry, tag column), explicit when-not (anything that weakens protection is refused and points at the dashboard), plus an authorization prerequisite ('Admin or owner only') and a workflow rule ('Show the user the diff before apply_config_change').

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