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

Annotations mark readOnlyHint=false and idempotentHint=false, which would otherwise be puzzling for a read-style tool; the description resolves this by disclosing that the cursor advances by default so each change is delivered once, and that peek/since override that. It also explains the empty-list semantics, adding genuine behavioral context beyond the annotations.

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?

Three sentences, front-loaded with the core purpose before qualifying behavior, and every clause carries load: scope, the re-browsing rationale, the empty-list signal, and the cursor modes. No filler or repetition of structured fields.

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 no output schema it covers a lot: what an empty result means and that results are delivered once. It still does not describe the shape of a change record or reference has_more/pagination outside the schema, so a small return-value gap remains.

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 clears that bar by explaining the intent behind peek ('to look without advancing') and since ('to replay'), giving the agent the purpose of the two most non-obvious parameters. It does not add anything for limit, watch_id, or client_token.

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 with explicit scope ('job changes detected since your last get_changes call, across all your watches or one watch'), so an agent immediately knows what it returns. It does not, however, name or contrast itself with sibling tools such as ack_changes, which shares cursor-advancing behavior, so the sibling-differentiation criterion for a 5 is unmet.

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?

Gives clear context for when this is the right call ('the cheap replacement for re-browsing careers pages') and explains the three operating modes: default advance, peek=true, and since=<cursor> replay. It stops short of naming an alternative tool or stating when not to use it, so it lands at a strong 4 rather than 5.

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