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

Annotations only declare readOnlyHint=true, so safety is already covered. The description adds genuine behavioral context: it discloses the output shape at a high level ('Each issue names its line') and, more importantly, the scope limitation that it analyzes only the supplied text — a constraint that changes how an agent must call it. It stops short of describing result fields, severity, or any contract-lookup dependency.

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 the action and resource, followed by the check inventory and then the usage constraint. The long enumerative middle sentence is dense but every clause names a distinct defect class, so it earns its length; only the sheer run-on length keeps it from a 5.

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 must stand in for return values, and 'Each issue names its line' plus the check inventory gives an agent a usable expectation of the response. Combined with the whole-file calling constraint and the readOnly annotation, nothing essential to invoking it correctly is missing, though result structure remains only loosely sketched.

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 0% — the single 'markup' property has no description at all — so the description must carry the meaning, and it largely does: it specifies that the value is a full page or file including <style> blocks and style attributes, and that excerpts produce partial results. It never states the expected type or format (string of HTML/JSX), but the semantics of what to pass are clear.

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 opens with a specific verb and resource ('Checks HTML or JSX') and then enumerates the exact defect classes it detects (nonexistent classes, variant misuse, hand-written colors, off-scale spacing, overridden tokens, opacity-faded text). Against a sibling set of pure get_* retrieval tools, it is unmistakably a validator, so an agent can route to it without opening the 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?

It states when to invoke ('Run it on markup you wrote before finishing') and how to call it correctly ('pass the whole page or file, <style> blocks included, not an excerpt'), including the reason ('it only checks what it is given'). No sibling is a real alternative for validation, so the absence of named alternatives is not a gap.

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