pixel_compile

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

Annotations already declare idempotent=true, destructive=false and readOnly=false, so the bar is lower. The description still adds real value: it discloses the output artifact (a contact sheet), that compilation persists an 'ordinary recipe revision', and precisely what the metadata flag surfaces (pose joints, contacts, timing, autotile mask table, particle counts, fluid figures, wave counts, normal directions). It stops short of failure modes or cost, but the additions are substantive.

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?

A single sprawling run-on sentence whose operative clause ('Compiles the source...') is buried past a long comma-separated catalogue of particles, spells and fluids. The enumeration is not front-loaded and much of it duplicates what the source schema already implies about document families. It is under-structured rather than merely long.

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?

With no output schema and a nested object parameter, the description does supply the return shape (contact sheet plus optional metadata) and points to pixel_help for document formats, which keeps it viable. But for a compile step it says nothing about validation behavior, when it fails, or whether output is rendered or just a sheet, leaving an agent uncertain about sequencing with pixel_validate and pixel_render.

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?

At 50% schema coverage the description must compensate, and it does for the undocumented boolean: it enumerates exactly what metadata=true returns for each source type. The source parameter is already well covered by the schema (document families, inline/file/revision forms), so the description need not repeat it. Minor gap: no per-source-type constraints on what a valid source must contain.

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 actual action ('compiles the source into an ordinary recipe revision and returns a contact sheet') appears only after four lines of asset-type enumeration, so the verb+resource is real but buried. It never distinguishes itself from siblings like pixel_validate, pixel_inspect or pixel_render, and the opening 'Build a character animation...' line frames it as a content generator rather than a compiler. An agent can eventually infer the purpose, but the noisy feature dump dilutes it.

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?

There is no statement of when to use this versus pixel_validate (check first?), pixel_inspect (what would it return?), or pixel_render (render output vs compiled revision). The only routing hint is a parenthetical 'see pixel_help topics' inside the schema, which points at documentation, not at a sibling tool choice. No prerequisites or exclusions are given.

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