Relendi — Commercial Real Estate Loan Tools

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

Annotations already mark the tool read-only and non-destructive, and the description adds meaningful behavioral detail: the waterfall ordering, inclusion of NY CEMA savings, and a critical consequence of omitting prepaymentPenaltyPercent. This goes beyond the annotation-only safety profile, though edge-case handling and output specifics are not covered.

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 dense sentences, front-loaded with the primary behavior, followed by an important caveat. No filler or repeated schema content; every sentence earns its place.

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 11-parameter financial tool with no output schema, this description gives a good high-level picture and the most critical input caveat, but it leaves gaps: exact dollar/percent units are largely implicit, the closing-cost source is not reflected in the schema, and the output return format is not specified. It is adequate for selection but not fully complete for invocation without relying on parameter names.

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?

With only 18% schema description coverage, the description partially compensates by explaining the prepaymentPenaltyPercent requirement and linking maxLtv, minDscr, and minDebtYield to the LTV/DSCR/debt-yield constraints, plus the NY CEMA condition. Several parameters like noi, propertyValue, and amortizationYears rely on self-evident names and are not explicitly defined.

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 names a specific operation ('cash-out refinance waterfall') and states the exact output: the largest new loan supported across LTV, DSCR, and debt-yield, then net cash after payoff, prepayment penalty, and closing costs. This clearly distinguishes it from siblings like size_loan or estimate_closing_costs.

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 phrase 'Full cash-out refinance waterfall' establishes a clear invocation context, and the NY CEMA remark adds a conditional use case. It does not explicitly name sibling alternatives or exclusion conditions, so it falls short of a full routing guide.

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?

The description adds meaningful behavioral context beyond the annotations: answers are grounded in the knowledge base, short, and include citations. It also implicitly communicates that the tool is not exact by routing loan math to exact tools. Annotations already cover the read-only and non-destructive safety profile, so the description does not need to repeat that.

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 two sentences with no wasted words. It front-loads the purpose and follows with a specific routing directive, avoiding repetition of schema or annotation content.

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 single-parameter read-only Q&A tool, the description covers the tool's purpose, input semantics, output nature (short, cited, grounded), and when to choose a sibling. No output schema is present, but the description adequately explains return characteristics, so the definition is complete for correct invocation.

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%, so the description must compensate. It defines the single parameter as a plain-English question about commercial real estate lending, adding semantic meaning beyond the bare string type and length constraints. For one parameter, this is sufficient, though it lacks examples or explicit topic boundaries.

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 clearly identifies the tool as a Q&A tool for commercial real estate lending, stating a specific verb, resource, and output (short answer grounded in Relendi's knowledge base with citations). It also distinguishes itself from exact calculation siblings by explicitly naming size_loan and run_scenario.

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 not to use the tool: for loan math, prefer size_loan or run_scenario, which are exact. The 'Ask a plain-English question' phrasing establishes the intended use case, and the exclusion draws a clear boundary with the computational siblings.

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 mark the tool read-only and non-destructive; the description adds genuinely useful behavior: metrics are computed only when their inputs are present, and each returned metric includes the formula used. This goes beyond what the annotations alone communicate.

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, front-loaded with the core purpose and metric list. No filler or restatement of the title beyond the essential.

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 multi-metric tool with ten optional inputs and no output schema, the description gives a good overview but does not specify output formatting, precedence rules (e.g., if netOperatingIncome and revenue-expenses are both supplied), or how inputs combine for each metric. It is adequate but not fully 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 schema already covers 60% of parameters with descriptions, and the remaining parameter names are self-explanatory. The description adds no new unit or relationship details beyond the optionality already visible in the schema, though it does reinforce that partial inputs are acceptable.

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 precise verb and resource ('Compute the standard commercial real estate underwriting metrics from raw figures') and enumerates the exact output set. This makes it easy to distinguish from siblings like calculate_payment or size_loan, even though it never names them.

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 clear context: call this when you have raw figures and need standard CRE ratios, and it explicitly invites partial input ('Supply whatever you have'). It lacks named alternatives or when-not-to-use conditions, so it stops just short of the highest tier.

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 destructiveHint=false, so the safety profile is known. The description adds value by specifying the computational outputs (balloon balance, total interest) and the relationship between term and amortization, which goes beyond the annotations. It does not contradict the annotations and adds meaningful 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.

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

The description is a single, well-structured sentence that front-loads the core purpose and enumerates the key outputs. It is concise with no filler, effectively conveying all essential information in a compact form.

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 fully explains what the tool returns (monthly and annual debt service, total interest, balloon balance) and gives the typical use case. This is complete for an agent to understand the tool's capabilities and call it correctly, given the read-only annotations.

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 40%, so only the rate parameter has a detailed description (DECIMAL). The tool description explains the concept of term vs. amortization and the balloon balance, which indirectly helps understand termYears and amortizationYears, but it does not explicitly map each parameter to its role. It provides some compensation for the low coverage but not complete guidance.

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 (calculate) and resource (payment math for commercial loans), and explicitly enumerates the outputs (monthly/annual debt service, total interest, balloon balance). It distinguishes itself from siblings like size_loan or calculate_metrics by focusing purely on payment math. The mention of the typical CRE case (10-year term on 30-year schedule) further clarifies its scope.

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 clearly indicates this is for loan payment calculations, which is a distinct use case from siblings like stress_test or compare_loan_offers. However, it does not explicitly name alternative tools or state when not to use it, only implying the context through the CRE example. It provides clear context but lacks explicit exclusions.

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 declare readOnlyHint true and destructiveHint false, covering safety. The description adds valuable context about the calculation methodology (accounting for origination, exit, and other fees, correcting misordering) and the output (effective rate, total cost, fee burden). This goes beyond annotations and gives insight into the tool's behavior.

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, no redundancy. The core purpose is front-loaded, followed by a brief explanation of the calculation logic. Every sentence earns its place, making it appropriately sized and well-structured.

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 specifies the output (effective rate, total cost, fee burden) which is essential given no output schema. It also explains the rationale, making the tool's behavior clear. It does not mention the minimum of 2 offers (which is in the schema) or edge cases, but for a comparison tool with moderate complexity, this is adequate. A 4 reflects the slight gap in explicit prerequisites.

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 includes descriptions for term and interestRate, but the overall schema description coverage is reported as 0%. The description mentions fee types (origination, exit, other fees) but does not explain individual parameters like loanAmount or the structure of the offers array. It provides some meaning about the overall logic but not enough to fully compensate for the low coverage. A 3 reflects the partial compensation.

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?

Description explicitly states the tool's function: ranking loan offers by true cost of capital, not headline rate. It specifies the resource (loan offers) and the action (compare/rank), and differentiates from siblings by focusing on true cost with fee adjustments. This is a specific verb+resource that distinguishes it clearly.

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 implies usage when you have competing loan offers and need to compare their true costs. It does not explicitly mention alternatives or when not to use it, but the purpose is clear enough that an agent would know to use it for comparison scenarios. Lacks explicit exclusions, so a 4 is appropriate.

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 mark the tool as read-only and non-destructive. The description adds useful behavioral context beyond that: it states the return shape ('every line item plus totals') and explains why NYC vs. upstate matters, including the existence of a tax that only applies in the five boroughs. Nothing contradicts the 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?

The description is two sentences with no filler. The first sentence front-loads the purpose and scope; the second delivers a critical input constraint with reasoning. Every clause earns its place.

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 the 12-parameter schema and no output schema, the description compensates well by indicating the output content and highlighting the most important conditional input (isNYC). It could be more complete by explaining how optional inputs like insurance or property tax affect the estimate, but the core invocation guidance is present and sufficient.

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 only 17%, so the description must compensate. It does clarify the state parameter by scoping to NY/FL and adds meaning to isNYC by mandating the NYC distinction serious. However, most numeric parameters (loanAmount, interestRate, annualInsurance, existingMortgageBalance, etc.) are left unexplained by the description and rely on their names and ranges for meaning, so compensation is incomplete.

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: 'Itemize commercial closing costs' for a New York or Florida transaction candidates. It lists concrete cost categories (lender, title, government, prepaid, third-party) and says the tool returns line items plus totals, clearly separating it from sibling analytics like calculate_payment or size_loan.

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 gives clear context for when the tool applies — estimating NY or FL commercial closing costs — and adds a mandatory usage condition for New York: the agent MUST specify whether the property is in NYC. It does not explicitly say 'use this instead of X' or describe when not to use it, so it stops short of a 5.

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?

The description clearly discloses that the tool does not save or submit anything, and that persistence only happens after the user follows the link and signs up or signs in to save a private draft. It also warns against including personal or account information. This substantially complements the readOnlyHint annotation and contradicts nothing.

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 four short sentences, each earning its place: what action is performed, when to use it, what data to include, and what happens afterward. It is front-loaded with the verb and resource and contains no filler.

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 15-parameter tool with no output schema, the description covers the essentials: the deliverable, the triggering user intent, data provenance rules, non-persistence, and privacy. It stops short of documenting nested structures or the exact form of the returned link, but an agent can 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.

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

With schema description coverage at only 7%, the description adds an important governing rule: include only facts the user supplied, and never put guessed values or calculated loan limits into requestedLoanAmount. However, it does not explain nested objects like address or assumptions, nor clarify units beyond what the sparse schema descriptions 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 names a specific deliverable ('a Move this deal to Relendi link') and ties it to the user's CRE scenario. Phrases like 'keep a deal, continue in Relendi, or carry property details beyond a calculator' clearly distinguish this from the analysis/computation-focused 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?

It explicitly states when to use the tool: when the user wants to keep a deal, continue in Relendi, or carry property details beyond a calculator. It does not name specific alternatives or exclusions, but the given context is sufficient to route an agent appropriately.

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 destructiveHint=false, covering safety. The description adds behavioral traits beyond that: it is 'deterministic' and uses 'same math as size_loan.' This clarifies output consistency and computational basis, which is valuable context not present in annotations. It does not describe return format or performance limits, but the deterministic note and math reference suffice for a strong score.

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 two sentences with the primary purpose front-loaded and a concise deterministic note at the end. It is efficient with no redundancy. It could be more structured (e.g., bullet points for modes), but it remains clear and readable.

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 8 parameters (including nested objects), no output schema, and only a high-level description, the tool is not fully specified. The description does not explain what the output looks like (e.g., a table or list of scenarios), nor does it detail the required inputs or how the modes differ operationally. For a complex tool like this, an agent would need additional guidance on expected inputs and outputs to invoke it 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 0%—the description mentions no parameters at all. The schema itself provides types and ranges but no semantic meaning (e.g., what 'rates' or 'thresholds' represent). The description only hints at the mode parameter via the three scenario types but does not explain required inputs (netOperatingIncome, estimatedValue) or how thresholds and NOI deltas are used. With 0% coverage, the description fails to compensate, leaving agents to infer parameter purpose from names alone.

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 clearly states the tool's purpose: running what-if loan sizing scenarios across assumptions (rate ladder, NOI sensitivity, loan feasibility). It explicitly distinguishes itself from size_loan by noting 'same math as size_loan, swept,' which differentiates it from the sibling tool. The verb 'Run' and resource 'loan sizing' are specific.

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 implies when to use this tool—when a range of assumptions is needed rather than a single calculation—by contrasting with size_loan. It names the alternative (size_loan) and notes the sweeping behavior, giving context for selection. However, it does not explicitly state when not to use it (e.g., for single-point calculations), which prevents a 5.

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 destructiveHint=false, so the safety profile is covered. The description adds useful behavioral detail beyond annotations: it returns 'source-cited excerpts' and operates on a curated knowledge base, which helps the agent set expectations for output and coverage.

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 two sentences with no filler. The core function and resource are front-loaded, followed by a concise output characteristic and a clear use case. Every sentence earns its place.

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 simple retrieval tool with annotations covering safety, the description is mostly complete: it names the knowledge base, the domain, the output format, and the intended use. However, the lack of any parameter-level explanation, combined with no output schema and 0% schema description coverage, leaves a gap in the call contract.

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%, so the description carries the burden of explaining parameters, but it does not. It never explains what 'query' should contain or what 'topK' controls, aside from the schema's min/max constraints. The agent can guess that topK means number of results, but the description adds no explicit parameter semantics.

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 ('Search') and a distinct resource (Relendi's curated commercial-real-estate knowledge base), and further specifies the domain content: lending concepts, deal lifecycle, fees, timelines. This clearly differentiates it from sibling tools like calculate_metrics or compare_loan_offers, which are analytical rather than retrieval 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 provides clear usage context: 'Use it to ground CRE answers rather than relying on memory.' This tells the agent when this tool is appropriate, though it does not explicitly name alternatives or exclusion cases.

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?

Beyond the readOnlyHint and destructiveHint annotations, the description discloses that the tool uses deterministic math, does not quote lender terms, and returns which constraint binds. This is meaningful behavioral context that helps the agent set expectations without calling 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?

Two focused sentences with no filler. The action, resource, calculation method, output, and an important exclusion are all packed efficiently, with the key result stated early.

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 tool with nested optional thresholds and no output schema, the description covers the main return contract and calculation scope well. It could note threshold defaults or the exact output shape, but the required inputs and core behavior are clear enough for correct invocation.

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 75% and the description adds value by explaining the four constraint categories that drive the calculation, which helps map the threshold fields like maxLtv, minDscr, and minDebtYield to their real-world meaning. It does not explain defaults or exact parameter relationships, but it goes beyond the raw schema.

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 clearly states a specific verb and resource: size a commercial real estate loan. It further specifies the unique output — maximum supportable loan as the minimum of LTV, DSCR, debt yield, and LTC — which distinguishes it from sibling tools like calculate_payment or compare_loan_offers.

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 gives clear context: use this tool when you need a first-principles, deterministic loan sizing calculation. It also gives an explicit exclusion by noting it does not quote lender terms, though it does not name sibling alternatives directly.

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?

The description adds behavioral context beyond the annotations by clarifying the output format (stressed DSCR per scenario with a verdict) and emphasizing the rate unit (percent). It does not contradict the readOnlyHint and destructiveHint annotations, and it provides useful detail about what the tool returns.

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 concise, with two sentences that front-load the core purpose and output. The clarification about rate units is essential and placed at the end without unnecessary filler. No wasted words.

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 omits important details about the stress scenarios (e.g., the magnitude of rate shocks or income decline, or whether they are predefined) and the role of minDscr in the verdict. Since there is no output schema, the description should provide a fuller picture of the expected behavior and inputs, but it leaves gaps that could confuse an agent.

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?

With only 33% schema description coverage, the description should compensate by explaining parameter usage. It only clarifies the rate unit (already in schema) but does not explain other parameters like netOperatingIncome, propertyValue, or especially minDscr, which determines the pass/fail threshold. The description does not mention how scenarios map to input 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 clearly states the tool's action ('Stress a loan against rate shocks, income decline, and combined scenarios') and specifies the output (stressed DSCR per scenario with pass/marginal/fail verdict). It is specific and understandable, though it does not explicitly differentiate from the sibling tool 'run_scenario', which could be a similar scenario analysis tool.

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 implies usage for stress testing (e.g., 'the same tests a credit committee runs'), but it does not explicitly state when to use this tool versus alternatives like 'run_scenario' or 'calculate_metrics'. There is no mention of when not to use it or any prerequisites.

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