Server Quality Checklist

Tool Scores

• edit_imageA
  3.7/5.0
  Behavior2/5

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

  Annotations are absent, so the description carries the full burden. It does disclose the default model and strength range, but it does not say whether the original image is overwritten, where the edited result is saved, or what the return value is. This is a notable gap for a mutation tool.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The purpose is front-loaded in one sentence, followed by a compact argument list. Every line adds value and there is no redundancy or filler.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness2/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  All parameters are documented, but the tool has no output schema and no annotations. The description never states what the tool returns or what side effects it has on the source file. For a 6-parameter tool with no other structured context, this leaves an agent unable to fully predict the tool's effect.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema description coverage is 0%, but the description compensates thoroughly. It defines all six parameters with useful detail: image_path path format and supported formats, prompt purpose, model default, size meaning, strength range and interpretation, and seed optionality. This goes well beyond the bare schema types.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description opens with 'Edit an existing image with a text prompt,' which states a specific verb, resource, and method. It clearly differentiates this tool from siblings like generate_image (create new) and text_to_speech (unrelated audio).

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The usage context is implied: edit an existing image. But the description never explicitly mentions alternatives such as generate_image or conditions when not to use this tool. The distinction from sibling tools is inferable but not stated.

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

• load_modelA
  4.1/5.0
  Behavior3/5

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

  With no annotations, the description carries the burden of behavioral disclosure. It explains the memory target and default-promotion side effect, and warns about chat-traffic implications, but it does not mention whether triggering a load can evict other models, whether it is idempotent, or any permission/error conditions.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is concise, front-loaded with the core purpose, and organized with an Args section. Every sentence adds useful information without filler.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness4/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a two-parameter tool with no output schema, the description is largely complete: parameters are precisely documented and routing/default behavior is clarified. It could go further by describing side effects on currently loaded models, but an agent can likely invoke the tool correctly with the information provided.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Despite 0% schema description coverage, the description fully documents both parameters: model is tied to list_models and server-side paths, and make_default is explained with concrete guidance about default promotion and side-loaded media models. This adds significant meaning beyond the bare schema.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description states a specific verb ('cold-load'), a specific resource (model on the remote mlx-serve server), and a specific destination (GPU memory). This clearly distinguishes it from siblings like list_models and unload_model.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description gives useful parameter-level context, especially the warning to leave make_default false for side-loaded media models. However, it does not explicitly say when to use this tool versus alternatives like unload_model or how it relates to automatic loading behavior.

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

• text_to_speechA
  4.1/5.0
  Behavior3/5

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

  With no annotations provided, the description carries full behavioral disclosure. It does mention that the tool returns a local WAV file path and provides speed range (0.25..4.0) and default values, which is useful. However, it doesn't disclose potential side effects (e.g., file cleanup, async behavior) or whether a model must be pre-loaded. It adds some context beyond the schema but not comprehensive detail.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is concise and well-structured: a one-line purpose, then a compact list of parameters with explanatory details. Each sentence earns its place, and the output format is front-loaded. No unnecessary repetition.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness4/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a simple tool with no output schema and no annotations, the description is nearly complete. It covers all parameters and the return value (WAV file path). It might benefit from noting whether the operation is synchronous or if any model loading is required, but these are minor gaps for a text-to-speech tool. Overall, it provides sufficient information for an agent to call it correctly.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema description coverage is 0%, so the description fully compensates by describing each parameter: text (what to speak), model (ID with default), voice (backend-specific), and speed (range and default). This adds significant meaning beyond the bare schema types and is essential for correct invocation.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description clearly states 'Synthesize speech from text; returns a local WAV file path,' which specifies a precise verb, resource, and output. It is easily distinguished from sibling tools (generate_music, generate_image, etc.) by the explicit text-to-speech purpose.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description does not explicitly mention when to use this tool over alternatives or provide exclusions. The distinct purpose (speech synthesis) implies usage, but there is no direct guidance such as 'Use this when you need spoken audio' or comparisons to other generation tools. The parameter defaults give some context but not usage conditions.

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

• unload_modelA
  4.4/5.0
  Behavior5/5

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

  No annotations are present, so the description carries the full burden. It goes well beyond the tool name by disclosing that unloading frees GPU memory, does not deregister the model, only evicts resident weights, and that the model can later be reloaded. This gives an agent an accurate mental model of the side effects.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is compact and front-loaded: purpose, behavioral consequence, and usage context appear in three tight sentences. Every sentence adds value and none repeat schema or annotation data.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness4/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a simple one-parameter unload operation, the description covers what happens, what remains after the operation, and when it is useful. It does not mention return values or error behavior, which could matter without an output schema, but the core information needed to select and invoke the tool is present.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters3/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  The input schema only defines a required 'model' string with no description, and schema description coverage is 0%. The description adds some semantics by implying the model is a registered model on the server, but it never specifies how the model identifier should be supplied or where it comes from. Given a single obvious parameter, the description partially compensates but does not fully document the expected value.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description starts with a specific verb and resource: 'Free a model's GPU memory on the remote mlx-serve server.' It clearly differentiates from siblings like load_model by explaining that the model stays registered and only resident weights are evicted. The behavior is unambiguous and not a tautology.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines4/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description gives a clear usage context: 'Useful after heavy media generation to reclaim unified memory.' It implies the relationship with load_model by noting the model can be reloaded later, but it does not explicitly name when not to use it or provide an alternative tool route. This is clear context without formal exclusions.

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

• generate_imageA
  4.5/5.0
  Behavior4/5

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

  With no annotations, the description carries the disclosure burden. It explicitly mentions that the tool saves a PNG file and returns its absolute path. It omits details like overwrite behavior or file lifetime, but the core side effect is visible.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is front-loaded with purpose and output behavior, followed by a compact and structured Args block. Every line adds value and the length is justified by the need to document seven parameters with no schema-level descriptions.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness4/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  Given that there is no output schema, the description correctly covers both the file-saving behavior and the returned absolute path. It also documents defaults and tuning parameters. It could be slightly more complete on routing to alternatives and file overwrite behavior, but an agent has enough to call the tool correctly.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema description coverage is 0%, so the description must compensate, and it does thoroughly. The Args section explains all seven parameters, including the model default, example sizes, meaning of seed, backend-specific steps, and the distinction between cfg_scale and guidance_scale.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  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: 'Generate an image from a text prompt on the remote server.' It also states the key output behavior (saved as PNG, returned path), which distinguishes it from siblings like text_to_speech, generate_video, and edit_image.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines4/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description clearly frames the text-to-image generation use case and points to list_models as the source for model ids. It does not explicitly say 'use edit_image instead for existing images,' but the context of generating from a text prompt is clear enough for correct tool selection.

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

• generate_3dA
  4.6/5.0
  Behavior4/5

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

  It discloses server-side execution, the image-compositing behavior, and returns a GLB file path openable in Blender/three.js/Quick Look. With no annotations, this carries the safety/behavior burden reasonably, though it leaves slight ambiguity about whether the default output is textured or untextured given the optional texture flag.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  Every sentence adds information: the first line is a crisp summary, then image requirements, return format, and a compact parameter list. It is long only because it carries type/range/default detail that no other structured field provides.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  With no output schema or annotations, the description is thorough enough to call correctly: input path syntax, defaults, ranges, prerequisite for texture, and output file type are all covered. No critical call-time detail is missing.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema description coverage is 0%, but the Args block documents all eight parameters with types, defaults, ranges, or conditions (e.g. octree_resolution [64,512] default 256; texture_steps [1,100] when texture=true). This fully compensates for the bare input schema.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  Opens with a specific verb+resource: 'Generate a textured 3D mesh (GLB) from a subject image on the remote server.' This clearly separates it from sibling tools like text_to_speech, generate_music, generate_video, and generate_image.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines4/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description gives concrete input prerequisites (clean cutout with alpha, opaque photos composited on white) and a hard condition for texture mode (paint weights must be installed, otherwise a named error). It does not explicitly name alternative tools or state when-not-to-use, but the 3D mesh target makes the decision context clear.

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

• generate_musicA
  4.6/5.0
  Behavior5/5

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

  No annotations are provided, so the description carries the full behavioral burden. It discloses the local-WAV side effect, the environment-variable model default, the output-length behavior for cover/complete tasks, the cover_strength range, and the backend-only caveat for steps.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The high-level summary is front-loaded, followed by a compact Args block. Each line maps to a parameter and adds unique information; there is no filler or repetition of the schema.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  Given the complexity (14 params, no annotations, no output schema), the description is complete enough to call the tool correctly: required prompt_style, optional parameters, task variants, constraints, and return type are all specified. Only minor operational details such as completion time or file cleanup are absent, and they are not essential for selection or invocation.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema description coverage is 0%, and the description compensates fully: every one of the 14 parameters receives meaning, defaults, ranges, examples, or accepted values (e.g. duration_seconds 10..600, task enum, vocal_language examples). The schema alone would be nearly useless without this.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The first sentence states a specific action ('Generate'), a resource ('a music track'), the input ('a style prompt'), and the output ('a local WAV path'). This clearly separates it from sibling generation tools such as generate_image, generate_video, and generate_3d.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  Use is implied by the music-generation focus and the task parameter values, but the description never explicitly says when to prefer generate_music over siblings like text_to_speech or generate_video. It provides strong internal task guidance (text2music vs cover vs complete), but no when-not-to-use statement.

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

• generate_videoA
  4.6/5.0
  Behavior5/5

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

  With no annotations, the description carries the full transparency burden and does so well. It reveals the ffmpeg/H.264 encoding step, local MP4 output, optional AAC audio, slowness, backend-specific frame ladders, default resolutions, and the 64-divisibility constraint for two-stage pipelines. This is substantial behavioral context.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is front-loaded with the core purpose and output, then briefly explains the pipeline and performance, and ends with a well-organized Args list. Despite covering many details, every sentence adds information and the structure makes it scannable.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a 12-parameter video generation tool with no annotations and no output schema, this description is remarkably complete. It covers output format, generation behavior, performance expectations, parameter rules, defaults, and optional inputs. An agent has enough to call it correctly.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters5/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema description coverage is 0%, but the Args section compensates fully for all 12 parameters. It gives defaults, backend-specific rules (LTX 8N+1, MiniMax-H3 17k+5), divisibility requirements, optional first/last frame conditioning, and audio duration matching. This is far beyond what the bare schema provides.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  Opens with a specific verb and resource: "Generate a short video from a text prompt; returns a local MP4 path." This clearly distinguishes it from sibling generation tools like generate_image or generate_music while stating the output format.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The intended use is obvious from the purpose statement and the media-specific wording, but the description never explicitly contrasts this tool with alternatives or states when not to use it. It also warns that generation is slow, which faintly implies a speed tradeoff, but no exclusions or alternative routing are given.

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

• health_checkA
  4.6/5.0
  Behavior4/5

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

  With no annotations, the description carries full behavioral responsibility. It discloses that the tool performs an HTTP GET to /health and is a non-mutating connectivity probe. It does not describe exact response semantics, but the purpose is clear enough for a health check.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  Three short sentences with no filler. The endpoint and primary purpose are front-loaded, followed by practical usage guidance. Every sentence contributes value.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness4/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a low-complexity, parameterless probe, the description covers what the tool does, how it does it (GET /health), and when to use it. It does not spell out the success/failure return payload, but the stated goal—reachability verification—makes the expected result reasonably clear even without an output schema.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters4/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  The tool has zero parameters and an empty schema with 100% coverage, so there is no parameter ambiguity. The description correctly adds no parameter details, and the baseline for zero-parameter tools applies.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description clearly states the action ('Check connectivity'), the target resource ('remote mlx-serve instance'), and the exact mechanism ('GET /health'). This is specific and easily distinguishes it from sibling generation and model-management tools.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines5/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description gives explicit guidance: 'Use this first to verify that the ip:port is reachable before running any generation.' This clearly tells an agent when to invoke it relative to other operations.

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

• list_modelsA
  4.7/5.0
  Behavior4/5

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

  No annotations are provided, so the description carries the burden of behavioral disclosure. It does this well by describing the return value: each model's id and capability flags for chat/vision, image, speech, music, video, and 3D mesh. The operation is clearly a read-only listing, and the default-model behavior adds useful context beyond the basic action.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is three sentences with no waste: it states the purpose, details the returned capabilities, and explains how the list is used by other tools. The most important information is front-loaded.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a zero-parameter listing tool with no output schema, the description fully covers what the agent needs: what the tool returns, which capabilities are reported, and how to integrate those ids into generation-tool calls. Nothing important is missing.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters4/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  The input schema is empty with zero parameters, so there is nothing to explain. The description appropriately focuses on output semantics rather than parameter mechanics, matching the baseline for a parameterless tool.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description uses a clear, specific verb ('List') and resource ('models available on the remote mlx-serve server') and expands on capabilities. It is clearly distinguishable from sibling generation and model-management tools.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines5/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The description explicitly tells the agent to pick model ids from this list for the model argument of generation tools, and clarifies that media tools without an explicit model use the server default. This is direct, actionable guidance for when to use the tool and how to use its results.

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