web_eval

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

No annotations are provided, so the description carries the full behavioral burden. It does disclose the return contract (last expression value, must be a plain JSON-serializable value), which is genuinely useful, but it omits session requirements (does it create or reuse a session?), sandbox/security restrictions, timeout behavior, and whether DOM or network access is permitted.

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 single sentence is well front-loaded with the verb and resource, but it is literally cut off mid-token, so the most operationally important clause (the required return type) is never completed. Size itself is fine; the defect is completeness of the sentence.

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 annotations, no output schema, no parameters, and a truncated description, an agent lacks the information needed to call this correctly: which session it targets, how the code is passed, and what shapes of return value are accepted. Only the return-value concept is addressed, and even that is incomplete.

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?

The schema declares zero properties, which normally sets a baseline of 4. But the description implies an implicit code argument ('Runs arbitrary JavaScript') and a target session, neither of which appears in the schema, leaving a real ambiguity about how the script is actually supplied. That unresolved mismatch pulls it below the zero-param baseline.

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 ('Runs arbitrary JavaScript in a web session') and names the return contract ('returns the last expression's value'), which is enough to distinguish it from siblings like web_extract or web_read. However, the sentence is truncated mid-word ('which must be a plain J...'), so the core value-type constraint is left incomplete, and there is no explicit sibling differentiation.

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?

No when-to-use guidance, no prerequisites, and no alternatives are named. Given siblings such as web_extract, web_find, and web_read, the agent gets no signal about when raw JS evaluation is preferable to those higher-level tools, or whether a web session must already exist.

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