direct_apply_changes

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

With no annotations provided, the description carries the full burden of behavioral disclosure. It clearly states that the operation is fail-closed, consumes the approval exactly once for idempotency, and performs a readback of affected campaigns to confirm the applied state. This is unusually transparent for a write operation.

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 two tight sentences with the core action front-loaded and the enforcement details following. Every clause adds necessary information—preconditions, idempotency, and readback—with no filler or repetition.

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 fail-closed write tool with no annotations, this description covers the critical operational context: what is required, what happens to the approval, and how the state is verified. The main gaps are the lack of explicit routing against direct_write and no explanation of the output beyond readback, though an output schema exists to cover return structure.

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?

The schema has 0% parameter descriptions, so the description must compensate. It does: 'accepted approval' maps to approval_id, 'consumed exactly once' maps to idempotency_key, 'affected campaigns are re-read' maps to campaigns, and 'allowlisted service/method' maps to service and method. It stops short of enumerating allowed values or formats, hence not a 5, but all five parameters receive meaningful semantic context.

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 specific verb and resource: 'Execute one accepted, allowlisted Direct write (Campaigns.update) with readback.' This precisely identifies the operation, the object being changed, and the verification step, and it clearly differentiates this tool from preparation or listing siblings.

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 invocation preconditions: a runtime flag, an accepted approval whose hash matches the exact changes, and a bounded allowlisted service/method. This tells an agent when it is safe to call the tool. It does not explicitly name alternatives like direct_write or direct_prepare_changes, but the 'accepted approval' framing strongly implies this is the post-approval execution path.

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