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

With no annotations, the description carries the burden and does well: it discloses that content is returned decrypted, that bytes arrive base64-encoded in data_base64, and that oversized files are silently unavailable through this path — a failure mode an agent must anticipate. It stops short of stating auth requirements or the error behavior when a name doesn't exist.

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 purpose and prerequisite, then the return shape, then the size caveat. The shell one-liner for decoding is practical but slightly instructional padding; the definition is otherwise tight for the amount of genuinely useful information conveyed.

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 read tool with an output schema, this covers purpose, prerequisite, return format, and the large-file escape hatch — everything needed to call it correctly. The remaining gaps (folder scoping, not-found behavior) are minor since the schema and output schema cover the structured side.

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 both 'name' and 'folder' are already documented in the schema. The description reinforces that the name must match what list_files reported but adds nothing about the optional folder parameter, so it neither compensates nor detracts. 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 and resource ('Retrieve one of the user's stored files, decrypted, by name') and explicitly separates the resource from a confusing sibling ('Files live separately from notes'). An agent can distinguish it from recall/remember 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?

Prescribes a prerequisite ('Call list_files first to find the name') and gives an explicit exclusion with an alternative ('Very large files are not inlined; retrieve those from the web console or the REST /file/download endpoint instead'). Both the when and the when-not are 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 and does so: it discloses irreversibility, a passphrase-verification prerequisite (an identity string alone cannot wipe), and the post-state ('no recall will return the folder's notes,' folder starts fresh). It even explains the underlying storage semantics — old encrypted notes persist on Walrus until expiry but are never served — which is unusually candid and prevents a misread of 'permanently'.

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 action and the confirmation requirement before any nuance. Roughly 110 words is defensible for a destructive, auth-gated tool, though the closing 'Honest semantics' paragraph is slightly discursive and could be tightened without losing substance.

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?

An output schema exists, so return values needn't be explained. Given a destructive, irreversible operation with a passphrase guard, the description covers confirmation, authorization, irreversibility, and post-state — everything needed to call it correctly and safely.

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 'folder' parameter's default-space behavior is documented in the schema itself, so the baseline is 3. The description adds no syntax, naming, or format guidance beyond what the schema already says about the folder argument.

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 first sentence names a specific verb ('Forget') and resource ('a folder of the user's memory') and states the effect ('permanently'). It reads clearly against siblings remember/recall/list_files/fetch_file, which are non-destructive operations on the same memory store.

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?

Explicit when-to-use and when-not-to-use: 'call it only when the user explicitly asks to forget or wipe a folder, never on your own judgement,' plus a required 'Confirm with the user first.' The gateway condition is stated unambiguously, which is exactly what an agent needs before invoking a destructive tool.

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 burden and does well: it discloses that files live in a separate store from notes, that results are name/size pairs, that an empty list is a valid non-error outcome, and that 'locked' true means the passphrase doesn't unlock them. It stops short of covering errors, permissions, or pagination behavior for large folders.

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 action and the return shape, then the routing hint, then edge cases. Slight redundancy: the 'files are separate from notes' fact appears in the description and again in the folder schema, and 'Call this to see which files exist' re-states the opening sentence.

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?

An output schema exists, so return values needn't be explained, yet the description summarizes them anyway and adds the semantic edge cases (empty list, locked) that the schema cannot convey. For a single-parameter read tool, an agent has everything needed to call it and interpret results.

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 schema already defines 'folder' with the default-space semantics, so the description adds essentially nothing new about the parameter beyond restating that files are separate from notes. Baseline 3 is appropriate where the schema does 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?

Opens with a specific verb and resource ('List the files the user has stored in a folder') and scopes the output ('name and size'). It also distinguishes itself from a sibling by stating that 'recall never returns them', which an agent cannot infer from the name alone.

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 sequences the workflow: 'Call this to see which files exist, then fetch_file to retrieve one.' It also rules out the misleading alternative ('recall does not find them'), covering both when-to-use and when-not.

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 behavioral burden and does so well: identity and passphrase come from MCP client configuration rather than arguments, decryption happens without exposing the secret, an empty list means nothing relevant is remembered, and `truncated: true` means the scan is incomplete. These details materially affect how an agent should call and interpret the tool.

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 front-loaded with the primary instruction to call at session start, then layers on re-use conditions, authentication context, and empty/truncated result semantics. Despite covering several behaviors, each sentence contributes useful operational guidance.

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 that an output schema exists, the description need not explain return structure in detail, yet it still explains the meaningful empty-list and truncation cases. Combined with authentication context and usage timing, it gives the agent everything needed to invoke and interpret this tool correctly.

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 already 100%, so both `query` and `folder` are documented structurally. The description reinforces the folder scoping and default-space behavior, but it largely repeats what the schema already states rather than adding new parameter-level syntax or constraints.

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 states a specific verb and resource: recall relevant information from the user's portable memory. It also distinguishes the tool from the sibling `remember` by explaining that recall should happen before remembering a digest so notes supersede rather than duplicate.

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?

It gives explicit timing and scope: call at the start of every session, scoped to the project's folder, and again whenever prior user context is needed. It also names a concrete alternative workflow, calling it before remembering a digest to avoid duplication.

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: it discloses that credentials come from MCP client config rather than arguments, that notes are encrypted per user and readable across devices, that writes land in a durable write-ahead buffer batched hours later, and that `stored: false` must be surfaced rather than assumed. It even pre-empts a likely misinterpretation by stating a `buffered:` receipt is healthy, not degraded — exactly the kind of behavioral context structured fields cannot supply.

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 purpose and trigger conditions, and every paragraph maps to a real agent failure mode (duplicate digests, misreporting buffered writes as failures). It is on the long side, and the explanation of Walrus per-write pricing is justification rather than instruction, but nothing is genuinely 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?

For a two-parameter write tool with an output schema, the description covers the remaining gaps an agent would hit: what `stored` and `receipt` mean, what to do on failure, and where credentials come from. No return-value documentation is missing that the output schema doesn't already provide.

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 baseline is 3; the description adds real meaning by clarifying that identity and passphrase are NOT parameters but come from client configuration, which prevents an agent from trying to supply them. It does not elaborate on `folder` organization beyond what the schema already states, keeping it just short of 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 and resource ('Store one thing in the user's private, portable memory on Walrus') and immediately constrains scope to one item per call. The contrast with 'recall' ('recall first so the digest supersedes earlier notes instead of duplicating them') lets an agent separate it from its closest sibling 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 concrete triggers — 'the moment something is worth carrying into a future session: a decision, a preference, a fact about the project, or an action taken on the user's behalf' — plus a session-end handoff-digest rule. It names the alternative ('recall first') and the ordering constraint, so when-to-use and when-to-prefer-recall are both explicit.

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