Create a runnable Astro website project folder with working instructions for an AI coding agent.
Scored across 4 tools
Three tools (get_starter, create_starter, start_website) all revolve around producing a ZIP, and the boundary between create_starter (creates a customized ZIP) and start_website (interviews then creates the ZIP) is genuinely blurry. describe_starter is clearly distinct, and the neutral-vs-customized distinction helps, but an agent could easily misselect among the ZIP-producing tools.
create_starter, describe_starter, and get_starter follow a clean verb_noun pattern with a shared suffix; start_website breaks the suffix convention but is still a readable verb_noun name. Minor deviation only.
Four tools is reasonably scoped for a guided website-starter generator, covering describe, download, guided workflow, and custom creation. It is on the thin side but each tool has a plausible role.
The surface covers discovery (describe), direct retrieval (get), guided creation (start_website), and custom creation (create_starter) with no persistent storage, so update/delete are not needed. Only a minor gap in post-creation continuation or validation.
Create a customized Website Starter ZIP after the start_website interview. Inputs are bounded and processed transiently; nothing is stored.
| Name | Required | Description | Default |
|---|---|---|---|
| aiTool | No | codex | |
| purpose | Yes | ||
| computer | No | private | |
| language | No | en | |
| projectName | Yes | ||
| dynamicNeeds | No | ||
| existingSite | No | ||
| publicLanguages | No | en |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The annotations already declare readOnlyHint, idempotentHint, non-destructive, and closed-world, so the safety profile is covered. The description adds one genuinely new behavioral fact – inputs are 'processed transiently; nothing is stored' – but 'inputs are bounded' is vague, and it says nothing about what the returned ZIP contains or how it is delivered.
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?
Two tight sentences with no filler; the action and the trigger are front-loaded before the data-handling note.
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 tool with no schema descriptions and no output schema, the description leaves all input semantics and the result artifact unexplained. It covers purpose and a privacy note but is far too thin for this parameter surface.
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 0% across 8 parameters, so the description carries the full burden and does not compensate. It never clarifies projectName, purpose, aiTool, computer, dynamicNeeds, existingSite, or publicLanguages, nor which two are required.
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?
Specific verb ('Create') plus a concrete resource ('customized Website Starter ZIP') and a sequencing constraint ('after the start_website interview') that cleanly separates it from the sibling start_website. An agent can distinguish it from describe_starter/get_starter without opening any 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?
States the precondition clearly: this is the post-interview step triggered after start_website, which is real when-to-use guidance. It does not name the other siblings (describe_starter, get_starter) as alternatives or state explicit exclusions, so it stops short of full routing guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Describe the Website Starter project, its current version, public download, and safety boundaries.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true, idempotentHint=true, destructiveHint=false, and openWorldHint=false, so the safety profile is fully covered structurally. The description adds the substantive content scope (version, download, safety boundaries), which is useful, but says nothing about caching, freshness, or what the response looks like.
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?
One tight sentence that front-loads the verb and resource and then lists the payload. No filler, though the enumerated aspects could be tightened slightly.
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 parameterless read-only introspection tool with no output schema, indicating what information is returned is the main requirement, and the description does that. The only gap is the lack of differentiation from get_starter.
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 tool takes zero parameters, so there is no argument semantics to explain and the baseline of 4 applies. The schema is trivially complete and the description correctly implies a no-argument invocation.
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 (Describe) and resource (Website Starter project) and enumerates the aspects covered: version, public download, safety boundaries. It does not, however, distinguish itself from the sibling get_starter, leaving an agent to guess whether 'describe' and 'get' return different things.
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?
There is no statement of when to prefer this over get_starter or create_starter, nor any prerequisite or exclusion. The agent must infer routing from the verb alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return a direct public link to the current neutral Website Starter ZIP. For a guided setup, use start_website first. No login is required.
| Name | Required | Description | Default |
|---|---|---|---|
| language | No | Language for the included instructions. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint, idempotentHint, and non-destructive, so safety profile is covered. The description adds meaningful context: no authentication required, and it returns a neutral default rather than customized output. This is genuine added value beyond annotations.
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 core action, then the alternative, then the precondition. 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?
Complete for a simple, single-optional-param retrieval tool with annotations and 100% schema coverage. Minor gap: it doesn't mention whether the link expires or how it's delivered, but this is a low-complexity read tool.
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% and the enum is documented in schema; description says nothing about the language parameter, so it adds no semantics beyond the schema. Baseline 3 applies.
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: 'Return a direct public link to the current neutral Website Starter ZIP.' Clear scope (neutral ZIP link) distinguishes it from siblings like start_website (guided setup) and create_starter.
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?
Explicitly names the alternative 'start_website' for guided setup and the condition that selects it. Also states 'No login is required', a useful precondition. Nothing left to inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Start the guided Website Starter workflow. Use this first to interview the person, create the ZIP, and continue in the generated folder.
| Name | Required | Description | Default |
|---|---|---|---|
| client | No | other | |
| language | No | en |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations declare readOnlyHint=true, yet the description explicitly says the tool will 'create the ZIP' and instructs continuing 'in the generated folder' — that is artifact/filesystem mutation, which directly conflicts with a read-only declaration. No additional context (permissions, idempotency, interview behavior) is provided beyond the conflicting claim.
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?
Two tight sentences with the routing cue and the workflow steps front-loaded; no filler. Slightly compressed to the point of under-specifying what actually happens during the interview.
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 and no parameter documentation, so the description must carry the full burden for a workflow-entry tool. It conveys the high-level sequence but omits what the interview collects, what the ZIP contains, and what to do with client/language — leaving an agent unable to call or follow up confidently.
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 0% and there are two enum parameters (client, language) with defaults. The description says nothing about either parameter — no meaning, no effect on the workflow, no guidance for the 'other'/'en' defaults — so it fails to compensate for the schema 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?
Names a specific verb ('Start') and resource ('guided Website Starter workflow') and states the sequencing role ('Use this first'), which separates it from create_starter/describe_starter/get_starter. It stops short of explicitly contrasting what the siblings do, but an agent can place it in the workflow.
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?
'Use this first to interview the person, create the ZIP, and continue in the generated folder' gives clear activation context and an entry-point role. It never names the alternative tools or states when NOT to use it, so it falls short of full when/when-not guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
start_websitecreate_starterdescribe_starterget_starterAI-native Git-based CMS for Astro. Create and publish content in the browser, without learning git.
Check how well AI agents can discover and understand a public website.
Generate and validate a .specs/ bundle for your repo, then hand it to your AI coding agent
Deploy static sites from AI agents: deploy_site publishes files and returns a live URL in seconds.
Add one secure layer between your agents and this server.
