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

Annotations only say readOnlyHint=false, openWorldHint=false, idempotentHint=false, destructiveHint=false, which tells the agent this is a mutating, non-idempotent, closed-world operation. The description goes far beyond that: it discloses default ducking behavior (-18dB bed vs -4dB lead), the context-aware switching logic, default loop behavior, default fades, offset_s semantics, and the smooth vs step ducking distinction. It also warns that storage_key must be an exact key and never invented. This is rich behavioral context that the annotations do not provide.

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 dense and information-rich, with front-loaded scope and a clear 'Mix music into the edit' statement. It uses capitalization and dashes to emphasize key terms (PROJECT-SCOPED, CONTEXT-AWARE, OUTPUT-timeline, SMOOTH). It is longer than a typical description, but every sentence adds operational value. It loses one point for being somewhat sprawling and for burying the 'Fades in/out by default' detail in the middle rather than grouping all defaults together.

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 an 11-parameter tool with no output schema, the description covers the critical decision points: how to choose music, how to handle defaults, how to override them, how to source storage_key, and how ducking behaves. It also references sibling tools for follow-up (set_music_fit) and evidence gathering (get_audio_analysis). The only minor gap is that it doesn't describe the return value, but the opening line says 'returns the project identity with its result,' which is sufficient for an agent to know what to expect.

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 18%, so the description must compensate. It does: it explains gain_db/duck as overrides, start/end as output-timeline seconds defaulting to the whole video, loop=true repeating a short track, offset_s starting partway into the track, and storage_key as an exact key from list_assets. However, it does not explicitly explain fade_in_s/fade_out_s or purpose beyond the schema's own description, and it doesn't enumerate all 11 parameters. Still, the most behaviorally important parameters are well covered.

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 clear project-scoped statement: 'this call acts only on the explicit project_id and returns the project identity with its result. Mix music into the edit.' It names the resource (music into the edit) and the verb (mix/add), and distinguishes itself from siblings like remove_music, swap_music, set_music_fit, and separate_music by focusing on adding music with ducking behavior. The description also clarifies scope (explicit project_id only), which differentiates it from tools that might use an active-project pointer.

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 when-to-use guidance: 'Choose from the user's direction, metadata, context and measured get_audio_analysis evidence.' It also tells the agent when to omit parameters ('omit them for 'add some music''), when to override defaults ('Pass gain_db/duck only to override that'), and how to source storage_key ('exact key from list_assets(kind='music')... never invent one'). It even references sibling tools (list_assets, get_audio_analysis, set_music_fit) for alternatives and follow-up actions. This is strong routing guidance.

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