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

Annotations only declare the operation is not read-only and open-world; the description goes far beyond by disclosing the async job_id model, the ~2 minute poll loop, that cancellation does not stop a running separation (credits are spent and output discarded), and that only plane-queued jobs get a failure refund. It also enumerates the exact result keys (stems, midi, drum_substems, vocals_lyrics, session_tracks, models_effective), so the agent knows what comes back despite there being no output schema.

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?

Purpose is front-loaded in the first clause and every subsequent clause carries operational information (async flow, result keys, cancel economics). It is dense and parenthetical-heavy — one long block rather than separated sections — which costs some scannability, but little of it is filler.

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 13-parameter, no-output-schema, expensive async job this is unusually complete: input sourcing (audio_key vs audio_url, 256 MB fetch limit), model selection, derivative outputs, result field inventory, polling path, and cancellation/refund behavior are all covered. An agent could call and follow up on this tool correctly from the description alone.

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 100%, so the baseline is 3, but the description adds cross-parameter meaning the schema only states piecemeal: stems_only implying include_midi:false, session_id requiring parent_track_id and yielding session_tracks/session_tempo_bpm, and guitar_split/other_classification outcomes flowing into named result fields. The optional idempotency_key semantics are also spelled out in behavior terms (same job returned after a timeout).

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 names a specific verb and resource ('split an uploaded mix into stems') and enumerates the outputs (vocals/drums/bass/guitar/piano/other or orchestral sets), plus a secondary framing ('vocal remover', acapella vs karaoke backing). That is enough to distinguish it from neighbors such as doseedo_split_soundtrack or doseedo_transcribe without opening either schema.

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?

It gives real context: use it to isolate vocals or produce karaoke backing, use stems_only for the fastest/cheapest run, use models=orchestra with instruments for wider sets, and poll with doseedo_wait_task. What it lacks is an explicit exclusion statement — nothing says when to prefer doseedo_split_soundtrack, doseedo_transcribe, or doseedo_session_from_stems instead.

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