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

Annotations mark readOnlyHint=false and idempotentHint=false; the description explains why by disclosing that the default call mutates cursor state while peek=true does not — a genuinely useful trait beyond the annotations. It stops short of covering auth expectations or error/rate-limit behavior.

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 tight sentences, front-loaded with the return semantics, then the empty-list signal, then the cursor/peek/since variations. No filler.

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?

With no output schema, the description does convey that the result is a list of changes and that an empty list is meaningful, plus the cursor advance contract. It omits pagination signals like has_more, which the schema mentions only on limit, leaving a small gap for a cursor-based paging tool.

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?

With 100% schema coverage the baseline is 3, but the description goes further by explaining how the default cursor position interacts with peek and since, which is meaning the schema alone does not convey. It adds no detail on limit or client_token beyond 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?

States a specific verb and resource — returns job changes detected since the last call, scoped to all watches or one watch. It also implicitly distinguishes itself from re-browsing careers pages and from sibling tools like search_jobs/watch_jobs.

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?

Explains the operational model clearly: default cursor advancement delivers each change once, peek=true looks without advancing, since=<cursor> replays, and an empty list means nothing changed. It does not, however, state when to prefer this over siblings such as search_jobs or how it relates to ack_changes.

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