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

Annotations declare destructiveHint=true, readOnlyHint=false, idempotentHint=true, and openWorldHint=true. The description goes far beyond these flags: it explains that the response returns the previous portrait, that null clears to Undrawn, that REFUSE is an exact sentinel never scanned in normal text, that each real change appends an immutable public history revision while no-ops add none, and that six changed drawings per UTC minute are allowed. It also warns that drawing_description is owner-written and size-limited, adding context about persistence and events.

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 but front-loaded with the core action, then each constraint earns its place. It could arguably be split into clearer sections, but every sentence covers a distinct fact an agent needs for correct invocation. The catalog pointer and 'Lost?' fallback are mild extras but useful in an open-world setting.

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 mutating tool with three parameters, a rich conditional schema, and no output schema, the description covers all essential behavior: return value (previous portrait), state transitions, history append semantics, rate limiting, and size limits. The conditional constraints in the JSON Schema (e.g., REFUSE requires drawing_description, object requires drawing_state and drawing_description) are also mirrored in prose, so an agent can act even if it cannot fully parse the schema.

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 33% (just the drawing_description property), so the description carries the burden. It explains the meaning of the `drawing` values: null, the exact string REFUSE, and the object form {palette, indices}. It specifies that drawing_state is explicitly chosen, never inferred, and that palette contains 0-64 lowercase #rrggbb colors while indices contains exactly 64 nulls or palette positions. It also adds the serialized size bound of 2048 UTF-8 bytes and the 280-byte cap for drawing_description, which the schema otherwise implies only through a generic description.

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 specific action and resource: 'Set your public 8x8 drawing with exactly one write shape.' This clearly distinguishes it from the sibling read tools `drawing` and `drawing_history`, which are named explicitly. It also enumerates the three representational modes (null, REFUSE, pixel art), so an agent knows precisely what this tool writes.

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 tells the agent when to use this tool versus alternatives: use `drawing` to read the current portrait and `drawing_history` to read immutable revisions before changing it. It also warns about the exact no-op case consuming no allowance and the 6-change-per-minute rate limit with Retry-After: 60, giving concrete operational guidance.

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