remove_background

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

With no annotations provided, the description carries the full burden of behavioral disclosure, and it does so extensively: it discloses upload behavior (local files and public HTTPS URLs uploaded automatically), auth requirements (MeiGen API key, MEIGEN_API_TOKEN or saved config), credit requirements (purchased credits only), retry semantics (at most once after suggested wait, preserve retryParameters), batch non-atomicity, and the fact that estimates are not spending caps. It also discloses that files are decoded, metadata removed, and GIF uses the first frame. The only minor gap is that it doesn't explicitly state the output format beyond 'transparent PNG cutout' and 'completed URLs', but the description is unusually transparent.

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 information-dense and front-loaded with the core purpose, but it is quite long and includes a large amount of operational policy (budget planning, recovery actions, caller ownership of progress) that, while valuable, could be trimmed or moved to a more structured location. Every sentence does earn its place in terms of content, but the sheer length makes it harder to parse quickly. It's not bloated with fluff, but it's at the edge of what an agent can absorb efficiently.

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 tool with 2 parameters, no output schema, and no annotations, the description is remarkably complete. It covers input requirements, auth, upload behavior, retry semantics, error handling, caller responsibilities, and interaction guidance. An agent has everything it needs to invoke this tool correctly and handle failures appropriately. The only thing not specified is the exact response shape, but the description says it returns 'structured status, handles, errors and completed URLs', which is sufficient given no output schema exists.

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 schema already documents both parameters well. The description adds meaningful context beyond the schema: it clarifies that productImage must be an actual source photo containing the subject, that relative paths are rejected, that files are decoded and metadata removed, and that GIF uses the first frame. It also reinforces the requestId reuse semantics. This goes beyond the baseline 3 by adding operational meaning to the parameters.

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 verb and resource: 'Create one transparent PNG cutout' for compositing, catalog assets, or logos. It explicitly distinguishes itself from scene creation ('it does not create a new scene'), which separates it from siblings like generate_ai_background and generate_marketing_poster. The purpose is unmistakable.

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: requires one actual source photo, no prompt or product facts needed, and it explicitly says to use the dedicated skill directly while skipping prompt enhancement, preference loading, and delegation. It also names the alternative path (list_skills for prices, check_skill for interruptions) and clarifies when to ask the user. This is comprehensive routing guidance.

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