computer_execute

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

Annotations already indicate readOnlyHint=false, destructiveHint=true, and openWorldHint=true. The description adds that it 'returns task and session IDs immediately' and that it 'remembers exact observed text for reuse,' which is useful. However, it doesn't detail the destructive aspects (e.g., what might be modified) or the open-world behavior beyond 'Jev chooses the sequence,' leaving some gaps that annotations only partially cover.

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 description is concise with five sentences, each adding value. It front-loads the core purpose and then covers key usage details. Some redundancy exists (e.g., 'Targets are available apps, not ordered steps' could be implied), but overall it's tight and well-structured.

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 tool's high complexity (nested objects, 8 parameters, no output schema) and zero schema coverage, the description covers the essentials: goal, until, targets, inputs, and session ID return. However, it omits parameter behaviors for maxSteps, timeoutMs, minConfidence, and ocr, and doesn't explain the openWorldHint implications (e.g., unexpected system changes). This is a substantial gap for an agent to call it correctly.

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 description coverage is 0%, so the description must compensate. It does explain some parameters: 'targets' are apps/browsers and browsers need URL, 'inputs' require exact text, and 'until' conditions are described (e.g., field_from_memory compares to remembered source). However, it doesn't explain 'ocr', 'maxSteps', 'timeoutMs', 'minConfidence', or the finer details of 'until' variants, leaving several parameters under-explained. It adds value over the schema but not enough for the coverage gap.

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 states the tool executes a goal across native apps and browsers, which is a specific verb+resource. It distinguishes from related tools like computer_act (which likely performs a single action) by focusing on multi-step goal execution. However, the term 'start one goal' is slightly abstract, and the description does not name specific siblings, relying on context to differentiate.

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 provides clear context for when to use this tool: when you have a goal that spans multiple apps or browsers, as opposed to single-step actions. It gives specific guidance on targets (browsers require URL, omitted targets discover apps) and mentions closing the session when finished, implying alternatives for single steps exist, though it doesn't explicitly name them.

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