Galley MCP

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

With no annotations provided, the description carries the full behavioral burden. It explicitly discloses safety ('never modifies the document'), the two possible compile backends (loopback service or browser live preview), and the output shape (errors/warnings + page count). This is strong disclosure beyond the schema.

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, each earning its place: what it does, how it does it, and when to invoke it. Front-loaded with the core action and result, with the usage guidance placed last as a natural follow-up.

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?

The description is complete for a zero-parameter, no-output-schema tool. It tells the agent the result format, the safety guarantee, the backend selection mechanism, and the precise workflow placement relative to siblings. No critical information is missing.

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 has zero parameters, so the baseline is 4. The description correctly avoids inventing parameter semantics and instead focuses on what the tool does and returns. Nothing is missing here.

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 uses a specific verb ('Type-checks') with a clear resource ('the shared project') and states what it returns ('diagnostics (errors/warnings + page count)'). It also clarifies a key boundary ('never modifies the document'), distinguishing it from the propose_* sibling tools.

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 explicitly says when to call the tool: 'Call this after propose_files/propose_edit to confirm the change set type-checks before asking the human to Accept.' This is clear context, though it does not spell out when not to use it or name direct alternatives.

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?

No annotations are provided, so the description carries the full burden. It does disclose the return payload ('pong' plus kernel version), which is the main behavioral fact, but it never states that the call is side-effect-free, whether it can fail/timeout, or how the optional echo parameter affects the response.

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 short sentences with zero waste, and the core fact (what it does, what it returns) is front-loaded. Nothing could be removed without losing information.

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 trivial no-output-schema liveness tool, the description covers purpose and return value adequately, but it omits any explanation of the one parameter and gives no indication of failure behavior. Acceptable but with a clear 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?

Schema description coverage is 0% and the description never mentions the single 'echo' parameter, leaving its purpose (likely echoing the input back alongside 'pong') entirely undocumented. With a non-zero parameter count, the baseline-4 zero-parameter rule does not apply, and the description adds nothing here.

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 operation ('Liveness check') and what it returns ('pong' with the kernel version), which is unambiguous against siblings that are all file/edit/compile operations. It does not explicitly name a sibling or contrast itself, but the function is distinct enough that no agent would confuse it with read_file or compile.

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?

'Liveness check' strongly implies when to call it (verifying the kernel is reachable), but there is no explicit guidance on when not to use it, no prerequisite notes, and no mention of alternatives. Usage is inferred 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?

With no annotations, the description carries the full burden of behavioral disclosure. It discloses size semantics (UTF-8 bytes when sizeExact, else a cheap lower bound), deterministic ordering, duplicate-path conflict flags, binary file pointer descriptors with exact sizeBytes, and the read-only nature of the operation. This is exceptionally transparent for a tool with no annotation support.

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 pack in the core action, output details, binary-file handling, and read-only context without redundancy. The most important information is front-loaded, and every clause adds value. The use of em-dashes and semicolons keeps related ideas together without bloating the text.

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 tool with no output schema, the description fully explains what the agent will receive: paths, sizes, duplicate flags, and binary pointers. It also explains how to proceed with binary files (pass to read_file) and clarifies the mutation boundary (propose_edit). Nothing needed to invoke or interpret the tool is missing.

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 has zero parameters, so the baseline is 4. The description does not need to explain parameters; it instead focuses on output semantics. It mentions 'sizeExact' as a field, not a parameter, which is appropriate. No parameter-related gap exists.

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: 'Lists the live (non-deleted) files of the shared project.' It then details exactly what is returned (path, size, deterministic order, duplicate-path flags) and distinguishes itself from read_file by explaining how binary file pointers are consumed by that sibling. This leaves no ambiguity about what the tool 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 provides clear context: it is a read-only listing operation, and edits are explicitly routed to propose_edit. It also tells the agent to pass binary file paths to read_file for the pointer. It does not explicitly compare against project_context or galley_ping, but the read-only/edit exclusion and the read_file handoff give enough practical guidance.

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?

With no annotations, the description carries the full burden and does so thoroughly: it declares read-only status, the budget-limited behavior, that duplicate-path conflicts and over-cap files are skipped and reported in `skipped` with reasons, and that `selectionTruncated` signals dropped chunks. This is unusually rich disclosure of what the caller will and won't get back.

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?

Front-loaded with the core purpose and provenance contract, then the budget/skip caveats; nearly every clause carries information. It is a dense single paragraph with nested parentheticals, which is slightly heavy for a two-parameter tool but not wasteful.

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?

There is no output schema, yet the description names the returned fields (file path, 1-based line range, heading path) and the two status signals (`skipped`, `selectionTruncated`) an agent needs to interpret a partial result. Nothing essential for correct invocation or result interpretation is missing.

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 baseline is 3, but the description adds meaning the schema lacks: it explains that budget caps excerpt text and connects it to the `selectionTruncated` signal, and clarifies that `query` ranks across the whole project rather than a single file. That is genuine added semantic value beyond the field descriptions.

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 ('Returns the most RELEVANT excerpts across ALL the project's live files for a query') plus the mechanism (lexical ranking over structural chunks) and scope (whole project, live files). An agent can immediately distinguish it from full-file readers like read_file.

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 routes to the sibling tool: 'Use read_file for a complete file,' giving a clear alternative with the condition that selects it. It does not spell out the inverse case (when not to use this tool at all, e.g., when you already know the file path), so it falls just short of a full when/when-not/alternatives statement.

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?

With no annotations provided, the description carries the full burden and does so well: it discloses the scratch-copy/publish model, that the document is never edited directly, the auto-accept possibility, the two possible status values, and the exact-match-once constraint with structured failure feedback. This is genuinely useful behavioral context an agent could not infer otherwise.

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?

Front-loaded with the core action, then behavior, then failure handling in a logical sequence with no filler. It is somewhat dense with parentheticals and quoted status values, but every sentence carries load.

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 two-parameter mutation tool with no annotations and no output schema, the description covers the mutation model, permissions/auto-accept nuance, expected return status values, and error recovery. Nothing critical for correct invocation is missing.

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 baseline is 3. The description goes beyond the schema by giving the `search` parameter operational meaning (must match exactly once, failures come back as status 'edit_failed' and should be refined and retried) and by noting the returned `status` must be inspected, which adds value over the static field 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 and resource ('applies search/replace edits to a scratch copy of the session's file and publishes the result as a proposal') and immediately distinguishes it from a direct-edit tool via 'It never edits the document directly.' An agent can tell what it does and that it is not the same as propose_files or read_document.

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?

Explains the two possible outcomes and when each occurs: auto-accept armed means status 'applied', otherwise a human Accept in Galley is required, and failures return status 'edit_failed' with a refine-and-retry instruction. It does not explicitly compare against sibling tools such as propose_files, but the usage context is clear.

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?

No annotations exist, and the description carries the full behavioral burden: it discloses that nothing changes directly, that auto-accept may apply automatically (status 'applied'), that the set is all-or-nothing, that edit search must match exactly once, that delete is a recoverable soft-delete, path safety rules, and the 32-op cap with no cross-proposal transaction. This is exactly the mutation semantics an agent needs and cannot get from schema 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?

Front-loaded with the atomic multi-file purpose and lifecycle before operational constraints, and nearly every sentence carries new information. It is a single dense paragraph, though, and would scan better split into purpose / lifecycle / op-rules / limits sections.

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?

There is no output schema, and the description compensates by explaining the return signal ('inspect the returned `status`') and how failures surface ('failures come back as structured data — refine and retry'). Combined with the lifecycle, path-safety, and size-limit coverage, nothing an agent needs to invoke this correctly is missing.

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 baseline is 3; the description earns above that by adding constraint semantics the schema doesn't state, notably the 32-op aggregate cap, the exact-match-once rule for `search`, and the distinction that create-binary-path reads a source file locally with no base64 transfer. It doesn't add per-parameter syntax beyond that, so not a 5.

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 with scope: 'Proposes a change set across MULTIPLE files as ONE atomic proposal', then enumerates the op kinds (create/edit/rename/delete/binary). The 'MULTIPLE' emphasis implicitly distinguishes it from the single-file sibling propose_edit, so an agent can route 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?

Gives clear when-to-use context (multi-file change sets), an explicit decision rule for binaries (inline base64 vs create-binary-path by absolute local path), and a concrete fallback for size limits ('split the work across multiple sequential propose_files calls'). It never names propose_edit as the single-file alternative, so routing to that sibling is left to inference.

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?

No annotations are provided, so the description carries the full behavioral burden. 'Returns' supports a read-only interpretation, and 'live shared project' hints at freshness, but the description does not explicitly state that the operation is non-mutating or describe failure behavior or prerequisites. It is adequate but minimal.

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?

A single, front-loaded sentence with no filler. Every word contributes to the meaning, and the structure is well suited to a zero-parameter read operation.

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?

Given zero parameters and no output schema, the description adequately explains what the agent will receive: the current text of the session's target file. It does leave the concept of 'session's target file' implicit and does not contrast with read_file, but for a simple read tool this is largely complete.

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 has zero parameters and the schema is empty, so there is no parameter semantics burden on the description. The description usefully clarifies that the implicit target is the session's target file, which is enough given the absence of 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 uses a specific verb ('Returns') and clearly identifies the resource (current text of the session's target file) and source (live shared project). It is clear in purpose, though it does not explicitly differentiate itself from the sibling read_file.

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 guidance on when to use this tool versus alternatives like read_file, nor any mention of prerequisites, exclusions, or context that would help an agent decide between tools. The description only states what the tool does.

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?

With no annotations, the description carries the full burden and does well: it discloses read-only behavior, a hard 2 MiB ceiling that errors rather than truncates, and the exact shape of the binary result ({ kind, path, hash, size, mime }). It omits what happens on a nonexistent or unreadable path and what the text response's structure is, which keeps it from a 5.

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 tight sentences, front-loaded with the primary behavior, then the refusal limit, then the binary exception. Every sentence carries a distinct operational fact and none repeats what the schema already says.

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 one-parameter, no-annotation, no-output-schema tool, the description covers the return payload in both text and binary cases plus the size limit, which is the information an agent cannot get elsewhere. Error semantics for missing paths and any permission requirements remain unaddressed.

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 single parameter is already documented, so the baseline is 3; the description still adds real meaning by forbidding normalization and globbing, which the schema's 'exact file path' phrasing does not spell out. That constraint changes how an agent constructs the argument, so it earns above baseline.

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 ('Returns the live replicated text of any project file') plus the exact-path constraint, which pins down scope precisely. Referencing list_files as the source of valid paths separates it from that sibling and from a generic document reader, so an agent can pick it out without opening the 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?

Usage is implied by the path requirement ('as listed by list_files'), which tells the agent to enumerate first, but there is no explicit when-to-use/when-not guidance against read_document, which is the closest sibling. The agent must infer the boundary between raw file read and document read.

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