hands_act

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

Adds genuinely non-schema information: consequential actions are gated by an approver, the channel can be switched to the phone via GHOST_HANDS_APPROVER, and the response reports which channel decided. That complements the destructive/openWorld/non-idempotent annotations with the governance model. It stops short of saying which kinds count as 'consequential' or what happens to state if an action is denied mid-flight.

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?

Two sentences, no filler, and the example payload is front-loaded where it is most needed. The approval clause is dense but each clause covers distinct ground; only the parenthetical channel switch could be tightened.

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?

For a destructive, open-world, non-idempotent tool with an opaque nested parameter and no output schema, the definition covers the approval gate and the response's decision-channel field but leaves the action-type vocabulary and the failure/rollback outcome unspecified. An agent still cannot enumerate what it may ask for without trial and error.

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% for a nested, free-form 'action' object, so the description carries the burden — and it does supply one worked example ({"kind":"click","target":3}) with the key field names kind and target. That is a meaningful lift over an empty object schema, but the full vocabulary of kinds and the meaning/source of target values remain undocumented.

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 ('perform one governed action ... on the perceived page') and gives a concrete example payload, so the agent knows exactly what operation this is. It differentiates itself from hands_run only implicitly through 'one action' and 'perceived page', never naming the sibling.

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?

'One governed action' and 'on the perceived page' imply this is for a single step taken after hands_perceive and not for multi-step flows (hands_run), but neither the alternative nor the ordering requirement is stated outright. The agent must infer routing from adjectives.

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