osd_perf

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

The '[READ]' marker and 'slowest first' ordering disclose real behavior beyond the readOnlyHint annotation, which is useful. However, nothing is said about what the output contains beyond latency metrics, whether results are truncated, or any scope limits.

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 short, front-loaded elements: a one-line purpose followed by the arg note. No filler, though the 'Args:' block formatting is slightly heavier than needed for a single optional parameter.

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 single-parameter read-only diagnostic with no output schema, the description gives purpose, ordering, and parameter default semantics — enough to call it correctly. Only the returned fields are left unspecified, which is a minor gap given no output schema exists to cover them.

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 carries the burden and does so: it explains that target is a Ceph target name sourced from config and that omitting it selects the default. This is meaningful semantic content the bare schema lacks.

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 resource and metric: per-OSD commit/apply latency, ordered slowest-first, with the intent 'find the slow disk'. It is distinguishable from the broadly-named sibling slow_ops, though it never names that sibling directly for contrast.

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 phrase 'find the slow disk' implies a diagnostic use case, which is more than nothing, but there is no explicit when-to-use vs. when-not guidance and no alternative tool referenced (e.g., slow_ops for cluster-wide slow op listing).

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