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

Annotations declare a non-readOnly, non-destructive, non-idempotent mutation, and the description adds real context beyond them: it is project-scoped only on the explicit project_id, the active-project pointer is never used to guess, it returns the project identity with the result, and 'auto' hue is solved immediately. It does not spell out reversibility or how the effect combines with an existing grade, but the disclosure is meaningfully richer than the annotations alone.

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 PROJECT-SCOPED constraint is front-loaded and each clause carries information about a distinct parameter or scope rule. It is dense with parenthetical mode explanations rather than wasteful, though the nested parentheticals make it slightly harder to scan than an ideal single-purpose sentence.

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 8-parameter effect tool with 38% schema coverage and no output schema, the description covers the scope contract, mode selection, and the default window, and notes the return of project identity. It omits the meaning of softness and tolerance, so an agent cannot fully reason about the effect's falloff, leaving the definition adequate but incomplete.

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?

With schema coverage at only 38%, the description compensates well for the key parameters: it defines keep='hue'/'subject'/'both', explains hue='auto' vs #rrggbb, clarifies keep_skin, and states the window defaults to the whole program for start/end. It leaves softness and tolerance entirely unexplained, which is the remaining gap.

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 gives a specific verb+effect: 'Keep one colour and send the rest of the picture to mono,' which clearly identifies a selective-colour isolation effect. It is distinct from siblings like match_grade or set_color_grade in effect terms, but it never names or contrasts a sibling, so it stops short of full differentiation.

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 keep/hue/keep_skin sentence implicitly tells the agent which mode to choose for a given intent, which is useful. However, there is no explicit when-to-use vs. when-not guidance relative to alternatives such as add_color_screen, match_grade, or add_subject_effect, and no prerequisites are stated outside the project_id scoping note.

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