MIL — Machine Intelligence Layer

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

Annotations only cover the safety profile (readOnlyHint, openWorldHint), so the description carries the response-contract burden — and it does: every requested identifier returns an explicit status (resolved, not_found, invalid_organisation_number) and absence is never signalled by omission. With no output schema this is genuinely useful, though the per-item error taxonomy beyond those three statuses is not elaborated.

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 sentences, zero filler, and the batch capability is front-loaded ahead of the result-contract detail. Every clause adds information an agent needs.

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 read-only batch resolver with a fully documented schema and no output schema, the description covers the essential unknown — how misses are reported. Minor gaps remain (no mention of partial-failure or ordering guarantees), but nothing critical 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 description coverage is 100%: the tier enum, pricing, x402 payment flow and the organisation_numbers pattern/limits are all documented in the schema. The description restates the 50-item cap but adds no format or semantic detail beyond it, so 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 ('Resolve'), resource ('companies') and batching scope ('up to 50 in one call'). The batch framing cleanly distinguishes it from the sibling get_company, so an agent can pick between them 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 batch size implicitly signals 'use when you have multiple identifiers', but the description never explicitly states when to prefer this over get_company or search_companies, nor any when-not condition. Usage is inferable 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 declare readOnlyHint=true and openWorldHint=true, so the description is not the sole source of the safety profile. It still adds material context annotations cannot: the free Basic tier vs paid tiers, the exact field set returned, and that sources/freshness are included. It does not cover auth requirements or failure behavior for an invalid organisationsnummer, which keeps it below 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?

Two sentences, front-loaded with the core action and input provenance, followed by the tier/cost caveat and its alternative. Nothing is repeated from the title or schema and every clause carries selection-relevant 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?

An output schema exists, so return-value documentation is not required, yet the description still lists the returned fields and flags cost tiers — enough for correct invocation on a one-parameter read tool. It omits only edge-case behavior (unknown or malformed organisationsnummer), which is minor.

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 100% and the single parameter is already documented as 'an id from search: the organisationsnummer'. The description restates the same meaning (id from search, organisationsnummer) rather than adding format, validation, or normalization details. Baseline 3 is appropriate when 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?

States a specific verb (fetch) and resource (register identity of one Swedish company), names the exact input source (the id returned by search, i.e. the organisationsnummer), and enumerates the returned fields. An agent can distinguish it from siblings like get_company or search 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?

Gives the precondition for use (id must come from search) and explicitly routes paid capability to an alternative: 'Verified and Intelligence (VAT/F-tax, history, annual reports, signals) are paid via get_company.' That is an explicit when-to-use-this vs when-to-use-something-else rule.

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 readOnlyHint=true and openWorldHint=true, so the safety profile is covered. The description adds meaningful behavior beyond that: signals are deterministic (no AI scoring), and each signal carries its rule, severity, supporting facts and detected_at — effectively documenting the response shape in the absence of an output 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 front-loaded sentences: what it finds, what a signal is, and what a signal contains. The example list is dense but informative rather than filler, and no sentence restates the tool name.

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 read-only, three-parameter tool with no output schema, the description covers purpose, signal semantics, and the fields returned, which is enough to invoke it correctly. Minor gaps remain: what "active" means, how many signals are returned, and whether filtering by severity/signal_type changes the result shape.

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 100%, and both optional filters (severity, signal_type) are fully enumerated in the schema, so the baseline is 3. The description echoes the severity values and gives a subset of signal-type examples, but adds no filtering syntax or semantics the schema does not already provide.

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+resource ("Find active warning signals for a company") and then defines the nature of the resource — deterministic rules over register facts and filed annual reports, not AI scores. The enumerated signal examples and the explicit mention of severity/rules/facts/detected_at make it clearly distinct from siblings like search_company_events or get_company_history.

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 "active warning signals for a company" but there is no explicit when-to-use guidance, no exclusions, and no pointer to an alternative such as search_company_events or get_company_history for raw events versus derived signals. The agent can infer the context but must guess at boundaries.

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 readOnlyHint=true and openWorldHint=true, so the safety profile is covered. The description adds genuinely useful behavioral detail: which relation types have backing data and that missing-data cases are reported with a reason instead of an empty list, which shapes how an agent should interpret results.

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 core behavior front-loaded; the second sentence handles the edge-case semantics. Slightly dense but every clause earns its place and nothing is wasted.

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 read-only, single-required-param lookup with no output schema, the description covers purpose, supported relation types, and the degraded-result behavior. Enough for an agent to call it correctly; only the shape of returned relation records is left implicit.

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 100%, so the parameters (organisation_number, limit, relation_type with enum) are already fully documented in the schema. The description reinforces the relation_type semantics but adds no new syntax, format, or default information beyond what the schema provides.

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 (find) plus resource (companies related to a given company) and enumerates the evidence-based relation types it covers (industry_cluster, address_cluster). This clearly distinguishes it from siblings like get_company and search_companies, which retrieve rather than traverse relations.

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 when each relation type is usable and that unsupported ones (board_member, ownership) yield an explicit unavailable status rather than an empty result. It gives clear context for calling the tool, though it does not name a sibling alternative to use for ownership/board-member queries.

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 only declare readOnlyHint=true and openWorldHint=true, so the description must carry the rest and does: it discloses that values carry confidence and source evidence, that freshness is reported, and that unknown data is returned explicitly rather than invented. That integrity guarantee is genuinely useful behavior an agent cannot infer from annotations. It omits any rate-limit or latency context (fair-use limits surface only in the schema's tier text).

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 tightly written sentences, front-loaded with the purpose, then return contents, then the data-integrity guarantee. No filler or restatement of the tool name.

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 compensates by enumerating the returned payload (attribute values, resolved knowledge, signals, events, provenance, freshness) and the confidence/evidence contract. Payment and tier behavior are thoroughly covered in the schema. Only minor gaps remain, such as response size or latency expectations for the intelligence tier.

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 100%: organisation_number's pattern/format and tier's enum values, pricing, and x402 payment retry flow are all fully documented in the schema. The description adds only the phrase 'by organisation number', which the schema already covers. Baseline 3 is appropriate when 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?

States a specific verb ('Retrieve'), resource ('machine-readable intelligence'), and access key ('by organisation number'), which clearly frames this as a single-company exact-identifier lookup. It never names a sibling, so the distinction from batch_get_companies, get_company_history, or search_companies is left to inference from 'by organisation number'. Clear, but no explicit sibling differentiation.

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 rather than stated: the phrase 'by organisation number' signals that this is the exact-identifier path versus the search_* tools, but no when-to-use/when-not guidance or named alternatives appear in the description. The tier selection logic that would be the most useful usage guidance lives in the schema, not the description.

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 only declare readOnlyHint and openWorldHint, so the description carries the interesting burden and does add real context: it discloses the event model (attribute_changed events, first-observed semantics, effective_from temporal bounds, evidence per event). Since there is no output schema, this return-shape disclosure is genuinely valuable. It still omits ordering and pagination behavior, and the presence of `limit` implies paging that is never explained.

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, zero filler, and the core capability is front-loaded before the optional filter and return-shape notes. Every clause carries 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 read-only 3-parameter tool with full schema coverage and no output schema, the description covers purpose, event semantics, filtering, and return content. The remaining gap is paging/ordering behavior implied by the limit parameter, which an agent calling this tool would want to know.

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 100%, so the schema already documents all three parameters, including the limit default/max and the organisation_number format. The description only restates the optional attribute_key filter and adds no syntax or semantics beyond it. Baseline 3 applies when 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?

States a specific verb and resource ('Retrieve the change history for a Swedish company') and narrows the content to 'attribute_changed events showing when values were first observed and what changed', which is far more precise than the bare name. It does not explicitly contrast itself with siblings like get_company or search_company_events, so the agent must infer the boundary, but the scope is unambiguous on its own.

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 pick this over get_company (current state) or search_company_events (event search). 'Optionally filter by attribute key' is parameter guidance, not tool-selection guidance. The agent is left to infer the use case entirely from the purpose sentence.

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 cover the safety profile (readOnlyHint, openWorldHint), and the description adds genuinely useful behavior beyond that: a hard 10-result cap, the presence of an id and citable URL per result, the free tier being limited to Basic identity, the refusal of signal-filtered queries, and name-search fallback for ambiguous input. It also clarifies that locality means registered locality, not nearby workplaces — a semantic trap otherwise invisible.

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 capability, and each sentence carries distinct information (formats, discovery examples, resolution rules, fallback, limits). It is dense as one long paragraph and would scan better as a short bulleted list of query forms, but there is little wasted 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?

An output schema exists, so return-value detail is not required, and the description still covers scope, input grammar, result cap, cost tier, and refusal routing. Nothing needed to invoke it correctly appears to be missing; only explicit pagination is absent, which is minor given the stated 10-result cap.

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% for the single query parameter, so baseline would be 3, but the description goes well beyond the schema by specifying org-number format (10 or 12 digits, dashes allowed), the SNI/alias resolution mechanics, and Swedish-language query support. This is substantive grammar the schema does not encode.

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 (search Swedish companies) and enumerates accepted input forms — name, organisationsnummer, and free-text discovery. It distinguishes itself from find_distress_signals by refusing signal-filtered discovery, but it never differentiates itself from the sibling tool search_companies, which an agent cannot disambiguate from this description 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?

Gives explicit when-not guidance ('Signal-filtered discovery is refused in free search; use paid find_distress_signals for a known company') and explains fallback behavior for unrecognized wording, plus names fetch as the follow-up consumer of returned ids. The remaining gap is the unaddressed overlap with search_companies.

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 readOnlyHint and openWorldHint, so safety is covered. The description adds real behavioral value beyond that: the return is 'compact summary cards' containing organisation_number, and the tool is positioned as the entry point of a two-step flow. It stops short of describing pagination or result-count behavior beyond the schema-level limit.

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-loads the core purpose and the non-obvious 'no organisation number needed' constraint before the filter list. The numbered discovery flow slightly repeats the get_company pointer already stated in the preceding sentence, which is minor redundancy in an otherwise efficient description.

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 carries the return-value burden and does so by describing compact summary cards and the organisation_number they contain. Combined with the stated optionality of all six parameters and the two-step discovery flow, an agent has everything needed to call it correctly and chain to get_company.

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 100%, so every parameter is already documented with its matching semantics (substring match, SNI prefix, locality combined with ANY matching SNI, has_signals tri-state). The description restates name/industry_code/has_signals filtering but adds no syntax or semantics beyond the schema, and omits locality and organisation_number entirely — baseline 3 is appropriate.

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 ('Search and discover Swedish companies') and immediately clarifies the key scoping fact: it works without an organisation number, which is exactly what separates it from get_company. The filter dimensions and return shape are also named, so an agent can distinguish it from siblings 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?

Provides an explicit discovery flow naming get_company as the follow-up alternative, and explicitly says all parameters are optional and can be omitted to browse everything. It does not address the other plausible siblings, such as find_distress_signals (which overlaps with the has_signals filter) or search, so the when-not guidance is incomplete.

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 readOnlyHint=true and openWorldHint=true, so safety is covered; the description adds value beyond them by disclosing the return payload (company identifiers, previous/new values, temporal bounds). It stops short of mentioning pagination or result-volume behavior beyond the limit parameter.

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 with the purpose front-loaded, then usage, then return shape. No filler, though the return sentence could arguably be folded in more economically.

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 compensates by describing the returned event fields. Combined with full schema coverage and read-only/open-world annotations, an agent has enough to call it correctly; only pagination/ordering details are 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 limit, event_type, and attribute_key are already documented in the schema. The description's "employee count drops" example loosely illustrates attribute_key semantics but adds no format or value guidance beyond the schema examples. Baseline 3 applies when 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?

States a specific verb+resource ("Search for company attribute change events") and scopes it "across companies," which visibly distinguishes it from single-company siblings like get_company_history and get_company. An agent can tell what it retrieves 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?

Gives two concrete use cases: finding companies where a specific attribute recently changed, and monitoring a dimension (with an example, "employee count drops"). It does not name an explicit alternative (e.g. get_company_history for a single company's timeline), so selection guidance is clear but not exhaustive.

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