audio_analyze

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

With no annotations, the description carries the full burden and does well: it discloses the side effect ("Writes a PNG of the plots next to the file") and the plot default, states the measurement noise floor, and explains the gating caveat that can hide real changes. What it does not state is whether it only reads (implied but never asserted), whether the file must be local, or whether other formats are rejected.

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 with the core action and no filler sentences, but it is essentially one very long run-on sentence packed with a comma-separated inventory of metrics. Since there is no output schema, that inventory largely earns its place; still, the density (band edge frequencies on every band, nested parentheticals) makes it harder to scan than it needs to be.

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 tool with no output schema, listing the returned measurements and the reference-difference behavior is exactly the right compensation, and the noise-tolerance note arms the agent to interpret results correctly. Remaining gaps are minor: it references an output field (`loudnessDeltaLu`) that appears nowhere in the input schema without saying so explicitly, and it never confirms the tool is read-only apart from the PNG write.

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 60%, and the description adds real meaning to the biggest gap: the `reference` parameter is explained as another WAV (a capture before a change or a commercial track) that yields a difference per measure, with the useful note that band differences are loudness-independent. `plot` is reinforced by "writes a PNG ... next to the file", and sections are implied by "loudness per section", though the `segment` default of 10 s is left entirely to the schema.

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 resource ("Measure a WAV") and then enumerates exactly what is measured: EBU R128 integrated/short-term/momentary loudness and LRA, sample/true peak, crest factor, PLR, per-band energy, low-end below 120 Hz, L/R correlation, stereo width, and section loudness. It also names its input sources (an audio_capture, a bounce, any mix), which cleanly separates it from the capture/bounce siblings that produce those files.

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 clear usage context: use a reference WAV when comparing before/after or against a commercial track, and use the ungated loudnessDeltaLu rather than integrated LUFS for short excerpts because integrated is gated for whole songs. It also warns that repeat captures differ by ~±0.2 dB true peak and ~±3% width, so smaller moves are noise — practical when-to-trust-it guidance. It stops short of explicitly naming an alternative tool or a when-not-to-use condition.

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