plan_capture

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

Beyond the readOnly/idempotent annotations, the description discloses that no measurement is run, that auto mode records the absence of enforced descendant containment, and that an adapter may still execute a bounded compatibility probe such as NVBench --version. The phrase 'runs the declared workload directly' is slightly ambiguous if read literally, but it is qualified by 'without running it' and 'never runs a measurement,' so there is no annotation contradiction.

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 compact and front-loaded: the core action appears in the first sentence, followed by the most important mode distinctions and the probe caveat. There is little wasted wording, though the indented line breaks add visual noise without structural value.

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 complex 10-parameter tool with nested option objects, the description gives useful mode-selection context and the no-measurement boundary but leaves important workflow context implicit: how a 'current capture' is obtained, when to supply each *_options block, and how the bound plan is meant to be consumed afterward. The presence of an output schema reduces the need to describe return values, but the surrounding lifecycle is still under-explained.

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 only 30%, and the description compensates only for capture_mode by explaining auto, managed, and trusted_local. It does not clarify adapter, workload_name, parameters, preflight_mode, external_context, or the profiler-specific *_options objects beyond their terse schema labels, leaving most of a 10-parameter tool semantically underspecified.

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 action and object: 'Bind one current capture without running it,' and reinforces the boundary with 'This tool never runs a measurement.' This clearly distinguishes planning/binding from execution-oriented siblings like execute_capture_plan or start_detached_capture. However, 'current capture' and 'bind' remain somewhat domain-specific and no sibling tool is named, so it stops short of full differentiation.

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 explicit conditional guidance for capture_mode sub-values: auto is the default, managed should only be used when project policy requires containment, and trusted_local requests direct local execution explicitly. It does not, however, say when plan_capture itself should be chosen over execution or other capture-planning siblings, and the prerequisite of a 'current capture' is only implied.

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