screen_do

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

With no annotations, the description carries the full burden, and it delivers: it discloses execution order, failure semantics ('If an action fails the run stops there and the error names it'), and the 5250 transmission model where typing sends nothing to the host until a key is pressed. This prevents a caller from assuming per-action round-trips.

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?

Every sentence earns its place: overview, batch rationale, action grammar, network-behavior caveat, and failure mode. The most important facts are front-loaded, and the examples make the syntax immediately understandable without padding.

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 still states what is returned ('only the final screen comes back') and what happens on failure. With no annotations, it explains the host-communication behavior and stopping semantics. Together this is enough for an agent to construct valid, correctly batched calls.

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?

Although schema coverage is 100%, the description adds a full action mini-language beyond the schema: valid prefixes like 'type', 'key', and 'cursor', accepted key names, field-addressing modes (ref, row,col, DDS name), and cursor behavior. It also clarifies the session parameter's default behavior and when to omit it.

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 concrete action and resource: 'Drive the screen: type into fields, move the cursor and press keys, in one call.' It clearly enumerates what the tool does and its one-call scope, and it distinguishes itself from observational siblings like screen_snapshot by being the interactive/input-driving tool.

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 gives explicit usage guidance: batch a whole interaction into one call rather than one action per call, with a rationale about order and the final screen. However, it does not explicitly name sibling alternatives such as screen_snapshot for read-only inspection, so the when-not-to-use guidance is implied rather than stated.

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