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

Annotations already declare readOnly/idempotent/non-destructive, so the bar is lower; the description still adds real context by naming the data source (official College Board distributions) and flagging that two of the four ranking metrics are only available for 2024 and 2025. It stops short of saying what happens if mean/test_takers is requested for 2026 or what the returned rows contain.

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 ranking operation and its metrics, then the invocation rule. The dense parenthetical about metric availability is slightly awkward and would read better as its own clause, but there is 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 5-param, all-optional ranking tool with no output schema, the description covers metrics, data source, defaults implied by schema, and the exam-list-vs-all behavior. Minor gaps remain: the row shape returned and handling of unrecognized exam names are unstated.

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 defines pass_rate as 'share scoring 3 or higher', explains the exams list purpose, and discloses the 2024/2025-only restriction on mean and test_takers, which is absent from the sort_by enum description.

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 (ranks) and resource (AP exams for one year) plus the exact ranking keys: pass rate, share of 5s, mean score, test takers. This is clearly a cross-exam ranking operation, distinguishable from siblings like get_ap_score_distribution (single exam distribution) and list_ap_exams (enumeration only).

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?

"Give a list of exams to compare a few, or leave it empty to rank all of them" gives explicit invocation guidance for the main parameter, plus a data-availability caveat (mean and test takers only for 2024/2025). It does not, however, name when to prefer a sibling tool such as get_ap_score_distribution for a single exam.

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 cover safety (readOnly, idempotent, closed-world), so the description's added value is the provenance and precision disclosure: exact counts for 2024/2025 but whole published percents for 2026, which tells the agent the 2026 figures are coarser. That is genuine behavioral context a caller needs in order to interpret results correctly.

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 filler: the first gives inputs and outputs, the second gives year-specific data provenance. Front-loaded with the resource and return values, so the agent gets the essentials first.

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 enumerating the returned fields and their meaning. Combined with annotations, an agent has what it needs; only failure behavior for an unrecognized exam name is left 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%, so the field descriptions already document exam aliases, the year range with its default of 2026, and the 1-5 score bound. The description mostly restates those bounds ('score from 1 to 5') rather than adding new syntax or edge-case meaning, so the 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 (estimate) plus the resource (percentile of an AP score) and enumerates exactly what comes back: percent lower/same/higher and midpoint percentile rank. It is clear enough to separate from list_ap_exams and compare_ap_exams, though it never explicitly contrasts itself with the closest sibling, get_ap_score_distribution.

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 framing 'For an AP exam, year and score from 1 to 5' implies the usage context, and the year note signals which years are valid. But there is no explicit when-to-use guidance and no naming of the alternative (get_ap_score_distribution) for callers who want raw counts rather than rankings.

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 readOnly/idempotent/non-destructive/closed-world, so safety is covered. The description adds real behavioral context beyond them: the data provenance (College Board's printed scoring worksheet and conversion chart), the guarantee that cut scores are never invented, and the error behavior for unsupported exams. It only omits how the error/edge cases look beyond the one mention.

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 action, then the supported-exam whitelist, then the anti-hallucination guarantee. The exam list is long but each item is load-bearing for correct invocation.

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 does describe the return values (a weighted composite and an estimated AP score), which is the key gap it needed to fill. Given only 3 parameters and closed-world reads, this is nearly complete, though it could say more about the shape of the returned composite.

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 a baseline of 3 is warranted, but the description adds meaning by explaining that MC and FR inputs are combined into a weighted composite, clarifying what the free-response array is for. It does not clarify per-exam input ranges beyond what the schema already documents.

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 transformation: multiple-choice correct answers and free-response points become a weighted composite and an estimated AP score. This is clearly distinct from the sibling tools (compare_ap_exams, estimate_ap_percentile, get_ap_score_distribution, list_ap_exams), none of which compute a score from raw responses.

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 an unusually explicit when-not clause: only the 7 listed exams are supported and 'other exams return an error', with each supported exam and year enumerated. The missing piece is a pointer to a sibling (e.g. list_ap_exams) as the way to discover supported exams, so it stops just short of naming 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?

Annotations already cover the safety profile (readOnly, idempotent, closed-world), so the description's real contribution is data provenance and coverage limits: sourced from College Board 2024/2025 PDFs and the 2026 distributions page, with mean/SD only 'where College Board published them'. That gap disclosure is useful behavior an agent cannot get from annotations. Return format details remain implicit.

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 return statement, then coverage notes. Three sentences, mostly earning their place; the 'newer than most model training data' clause is slightly promotional but conveys a real freshness caveat.

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 does the work of describing the payload (percentages, student counts, mean, SD) and the year coverage, which is adequate for a simple two-parameter read tool. It stops short of stating what happens for exams/years with no published data, which is the one gap an agent might hit.

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 documents the short-name examples for 'exam' and the 2024/2025/2026 availability plus default of 2026 for 'year'. The description adds little parameter-level meaning beyond restating that both inputs exist, so the 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: 'Returns the College Board score distribution ... for an AP exam and year', and enumerates what is included (5/4/3/2/1 percentages, 3-or-higher, counts, mean, SD). It is clearly a single-exam retrieval tool, but it never names or contrasts the siblings (compare_ap_exams, estimate_ap_percentile), so differentiation must be inferred.

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 only implied: the note that 2026 results are 'newer than most model training data' hints that this is the tool for current-year figures, and the year range is stated. There is no explicit when-to-use/when-not guidance and no pointer to the sibling tools for percentile estimation or exam comparison.

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, idempotentHint, and openWorldHint=false, so the safety profile is fully covered. The description adds genuine value by disclosing data provenance (source of each year) and the presence of composite-score worksheets, but says nothing about result size, ordering, or how the year parameter scopes the listing.

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 sentence with the main idea ('Lists every AP exam') front-loaded, followed by the fields it reports. Every clause carries information, though the trailing enumeration is dense enough that it reads as a run-on list of attributes.

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 spells out the returned fields (exams, years, per-year source, worksheet availability), which is the right compensation. The remaining gap is the interaction between the optional `year` filter and the 'years available for each' content, which an agent must guess.

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 lone `year` parameter is thoroughly documented there, including the allowed 2024-2026 range and the 2026 default. The description contributes nothing about the parameter and does not clarify whether passing a year narrows the list or the listing always enumerates every available year.

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 ('Lists') and a precisely scoped resource ('every AP exam with College Board score distributions in this server'), and even enumerates what each entry contains — years, per-year source, and worksheet availability. This reads clearly as a catalog/discovery tool, but it never names a sibling to contrast with, e.g. that per-exam detail lives in get_ap_score_distribution.

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 only implied: by describing a complete enumeration of exams and years, an agent can infer this is the discovery step before calling get_ap_score_distribution or estimate_ap_score_from_composite. There is no explicit 'use this when you need to know which exams/years exist' or any exclusion pointing at the estimation or comparison siblings.

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