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

Adds real behavioral context beyond annotations: originals are preserved, output goes to a new absolute path, and the call returns a job ID requiring polling via get_job. Annotations already cover readOnly/destructive/idempotent, so the description's job is supplementary and it delivers async-job and non-overwrite behavior.

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?

Four tightly packed sentences, each carrying a distinct fact (API used, output/preservation behavior, mask note, polling instruction). Fully front-loaded with no waste.

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 async, 9-parameter, open-world image-edit tool with no output schema, the description covers the key agent needs: it is async, results come via job ID, originals are safe, and where output lands. It could mention cost/charge implications (n increases charges, per schema) but the essentials are present.

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 78%, above the threshold where the schema does the heavy lifting. The description adds only a hint about mask semantics, not compensating detail beyond the schema's already-rich parameter docs.

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?

States a specific verb (Edit) and resource (local images), names the underlying API used, and distinguishes itself from generate_image by the editing/mask semantics. An agent can tell precisely what this does.

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 routes the agent forward: 'Poll returned job ID with get_job' tells the agent what to do after invocation, and mentions optional mask usage. It does not explicitly contrast with generate_image (when to edit vs generate), which is the main gap.

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

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

Annotations already declare readOnly=false, idempotent=false, openWorld=true and destructive=false, so the safety profile is covered. The description adds genuinely new context beyond that: the request is paid, retries are not automatic after timeout/failure, and 'existing files are protected'. It does not cover auth/permission requirements, but the marginal behavioral value is solid.

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?

Four short sentences, front-loaded with the action and cost implication, then output behavior, then the retry rule and file-protection guarantee. Nothing is padded or redundant.

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, the description helpfully characterizes the return ('paths/metadata or a job ID to poll with get_job') and covers cost, non-resubmission and file safety. For a 7-parameter paid mutation tool this is close to complete; the only gap is explicit prerequisite/alternative routing relative to edit_image and list_models.

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 71%, so the schema already documents model, prompt, output_path, n, size and quality in meaningful detail (format constraints, charge implications, provider defaults). The description adds no parameter-level syntax or defaults of its own. Baseline 3 is appropriate when the schema is doing the heavy lifting.

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?

States a specific verb and resource: 'Submit one paid image-generation request and save bytes locally', which an agent can distinguish from a sibling like edit_image by the generation-vs-edit framing. It also routes polling to get_job. It stops short of explicitly naming the alternatives it is not (edit_image, list_models), so it is clear but not fully sibling-differentiating.

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?

'Returns paths/metadata or a job ID to poll with get_job' gives real routing guidance, and 'Never automatically resubmit after timeout/failure' bounds retry behavior. However, there is no explicit statement of when to choose this over edit_image, nor any precondition (e.g., model must first come from list_models). Usage is implied rather than stated.

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

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

Annotations already establish the safety profile (readOnlyHint=true, destructiveHint=false, openWorldHint=false), so the bar is lower. The description adds genuine non-obvious behavior: jobs persist across MCP disconnects while the local worker lives, and interrupted jobs may still have usable outputs on disk. It omits what an invalid or expired job_id yields and how status is reported.

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?

Three sentences with the purpose front-loaded and no filler; each sentence carries a distinct payload (scope, durability, failure handling). Minor cost: the durability and failure-handling sentences blend behavioral fact with procedural advice, which slightly muddies the structure.

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?

No output schema exists, so the description is the only source for return shape, yet it never enumerates what a job query returns (status values, output paths, error fields) beyond alluding to 'output paths/provider logs'. For a two-parameter read tool with annotations covering safety, the coverage is adequate but leaves the agent guessing at response interpretation.

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 exactly 50%: wait_seconds carries a full description, job_id carries none, and the tool description adds no parameter guidance for either. The undocumented parameter is a self-evident required identifier and the description's 'existing image job' framing gives it meaning, so this lands at the baseline rather than below it.

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?

States a specific verb+resource ('Query an existing image job') and immediately bounds the scope with 'without submitting or retrying', which implicitly separates it from the submit-style siblings generate_image and edit_image. It stops short of naming those siblings, so an agent must infer the contrast rather than being told 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?

Gives clear when-to-use context (polling an already-submitted job) and a concrete decision rule for the hard case: on interrupted/unknown outcome, inspect output paths/provider logs 'before considering another paid request'. It does not explicitly say 'use generate_image to start a job', but the exclusion of submitting/retrying covers the alternative in practice.

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

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

Annotations already cover the safety profile (readOnlyHint, destructiveHint=false, openWorldHint), so the bar is lower, and the description adds real context: results are scoped by the configured API key, listing models does not establish generation entitlement, and image_candidate is only a name heuristic rather than a verified capability.

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?

Three short sentences, front-loaded with the tool's action and scope, with no redundant restatement of the name. Slightly telegraphic in the second and third sentences, but each contributes a distinct caveat.

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, the description usefully discloses that returned models carry an image_candidate field and that it is a name-based heuristic only. Scope, auth dependency, and the generation caveat are covered; pagination or result-size behavior is not mentioned, a minor gap.

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?

The single filter parameter is fully documented in the schema (100% coverage, including case-insensitivity and the omit-to-see-all behavior), so the schema carries the meaning. The description adds no syntax or format detail about filtering; its only field-level remark concerns image_candidate, which is a returned value, not an input.

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?

States a specific verb and resource ('List models') plus the scope that governs the result set ('visible to the configured provider API key'). It explicitly distinguishes itself from the sibling generate_image by stating it 'does not generate images,' so an agent can separate discovery from generation without opening either schema.

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 negative clause 'does not generate images or prove generation permission' tells the agent when NOT to reach for this tool and points the task toward generate_image. There is no positive framing such as 'use this first to discover model IDs before calling generate_image,' so the routing is inferable rather than explicit.

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