act

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

Goes well beyond the destructiveHint/openWorldHint annotations: batch halts on ~ or ✗ with nothing more sent and no replay, ambiguous labels are refused, outward actions need `confirm: true` plus an `accord` token from a prior refusal, and local destruction needs `confirm: true` alone. The confirm-vs-accord split and the 'no replay after unknown outcome' rule are exactly the operational facts an agent needs to avoid causing damage.

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?

Front-loaded and information-dense, but written in a compressed shorthand that costs parseability: 'Steps ✓ = ~ ✗, status = the worst' and 'follow `next`' require inference, and the action list runs as an unpunctuated stream. Nothing is redundant, but several lines do not earn their place because they cannot be decoded on first read.

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 non-idempotent, open-world actuator with no output schema and no schema descriptions, this covers safety gating, failure semantics, and targeting comprehensively. The remaining gap is the return value: step statuses and `next` are alluded to but never characterized, and three parameters are left dangling.

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?

With 0% schema description coverage the description must carry all parameter meaning. It covers hwnd, steps (with a full action vocabulary and example), accord, and foreground well, but `then`, `repeat`, and `time_budget` receive no explanation at all, and the referenced `next` is not even a declared parameter. Partial compensation only.

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?

Opens with a specific verb and scope: 'Act in ONE batch, in the background, on a NAMED window (hwnd), without reading the code.' That phrasing directly separates it from `look` (which reads the UI) and gives an agent a one-line mental model of what the tool is for.

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?

Gives the essentials for choosing this tool: batch-only, hwnd-targeted, `foreground=True` reserved as a fallback, and a hard precondition that a batch without a prior look is refused. It stops short of naming `look` explicitly as the required prior step, so the alternative 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.