splithub_results

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

With no annotations, the description fully carries the behavioral burden. It explains that verdicts are sequential (always-valid, safe to check any time), that guardrail_breached means do not ship, that sample quality never changes the decision, that stars are not a reason to override the verdict, that plain pValue is display-only while sequential pValue should be used, and that slices and views carry no voice or advice. This is unusually thorough for a results tool.

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 content is dense and almost entirely front-loaded into a single unbroken block with no paragraph breaks or bullet structure. While many sentences carry non-trivial information, the sheer length and run-on style make it hard to parse quickly, and the sheer volume reduces readability for an agent scanning for the most relevant usage rule.

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?

Given the complexity—an experiment-results tool with slices, views, self-hosted fallback, voice/advice distinctions, and a rich output schema—the description covers all critical decision points: how to handle missing analysis, how to treat guardrail breaches, when to invoke splithub_ship, and how to interpret sequential vs plain p-values. Despite the dense formatting, no essential behavioral or interpretative guidance appears missing.

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 schema already documents parameters, making 3 the baseline. The description goes further: it clarifies that from/to refer to first-exposure days in the project time zone, that filters keep visitors by country/device/platform/app_version/source/etc., that timeline means 'a day = visitors first seen that day, conversions whenever they happened', and that engagement returns property null when no analytics exist. These additions meaningfully exceed the schema text.

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 opening sentence 'Results of one experiment' plus the detailed breakdown of verdict fields makes the tool's function unmistakable: it returns the full statistical verdict for a single experiment. However, it never explicitly names sibling tools like splithub_experiment_status or splithub_experiment_events to clarify why an agent would choose this over those, so it falls short of the top score.

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?

The description gives concrete usage guidance: when analysis.available is false, report raw numbers and never call a winner; when finished, ask which variant to keep and then use splithub_ship; for slices, views, and segments, never call a winner from them. It lacks explicit 'use this instead of X' comparisons with sibling result-related tools, but the when-to-use and when-not-to-use conditions are otherwise richly specified.

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