Menso

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

Goes well beyond the readOnly/openWorld annotations by disclosing plan-tier behavior differences (Studio vs Free/Pro output), that returned text originates from the tested site, that it is wrapped in <site-content> tags, and a direct instruction to treat it as untrusted evidence rather than commands. This is exactly the kind of prompt-injection and data-provenance context annotations cannot carry.

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 what is returned before the plan-tier and safety caveats, and every clause is substantive. The single long sentence is dense but not padded; it could be split for readability, which is the only deduction.

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?

Output schema exists so return values need no further explanation, yet the description still clarifies output structure and plan gating, and it fully covers the one input. Nothing an agent needs to call this correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

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

Schema coverage is 100% and the single test_id parameter is fully documented (format, source from run_test), so the schema already carries the semantics. The description adds no format or sourcing detail beyond what the schema states, matching the baseline 3.

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 (get) and resource (results of a finished Menso test) and enumerates what is returned: TRACES score, per-dimension scores and reasons, task outcome, and plan-dependent friction points. This clearly separates it from siblings like run_test and get_status.

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 'results of a finished Menso test' implies this is only for completed tests, routing the agent to get_status for in-progress checks. However, it never names an alternative tool or states the precondition explicitly, so the guidance is contextual rather than spelled out.

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?

Goes well beyond annotations by disclosing the consequential side effect: the link is public, anyone can watch it, and it should be shared deliberately. It also explains the plan-based fallback behavior (upgrade note instead of a link). This is consistent with readOnlyHint=false and idempotentHint=true rather than contradicting them.

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 action, then the sharing warning, then the plan caveat. Every sentence adds decision-relevant information with no padding.

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 an output schema present, return values need not be described; the description instead covers the things structured fields cannot: public exposure, sharing caution, and plan-dependent outcomes. Nothing needed to invoke correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

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

Schema coverage is 100% for the single test_id parameter, including its pattern and provenance (returned by run_test), so the schema carries the parameter burden. The description adds no syntax or format detail, so 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 ('Create, or return the existing, public replay link for a finished Menso test') and pins the scope to finished tests. An agent can distinguish this from run_test, get_status, and get_findings 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?

The precondition (a finished test) and the plan gating (Pro/Studio gets a link, Free gets an upgrade note) tell the agent when the call is appropriate and what to expect otherwise. No sibling alternative is named, but the siblings are unrelated operations, so no explicit exclusion is needed.

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 (readOnlyHint=true, openWorldHint=false), and the description adds far more: the full state machine, the meaning of each state, the human-in-the-loop pause condition, and the recommended polling cadence. This is exactly the kind of behavioral context that cannot be inferred from structured fields.

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?

Purpose and prerequisite are front-loaded, followed by the state enumeration that gives each state an interpreting gloss, and it closes with the polling instruction. No sentence is filler despite the density.

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 need not be re-explained; the description instead supplies the interpretation of each state and the next action for terminal states. Nothing an agent needs to call and act on this tool is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

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

Schema description coverage is 100% and the single test_id parameter is already documented as 'returned by run_test'. The description's mention of run_test adds no syntax or format detail beyond the schema, 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+resource ('Check a Menso test') and ties it back to its sibling run_test, which produced the test_id. An agent can immediately tell this is the polling companion to run_test rather than get_findings or get_replay_link.

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 says to poll every 30-60 seconds and routes the agent to get_findings once state is 'done'. It also flags that 'paused' requires human action in the Menso web app, which is an actionable when-to-act condition.

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 cover the generic safety profile (readOnlyHint=false, openWorldHint=true, idempotentHint=false). The description goes well beyond them: credit costs (40 for speed, 300 for quality), the fact that the balance is checked before anything starts, expected duration of several minutes, and the side effect that 'signup' actually creates a new account. These are exactly the operational traits an agent needs before committing to a paid, non-idempotent run.

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?

Purpose is front-loaded in the first clause, and every subsequent sentence carries operational information (costs, template semantics, URL constraint, polling cadence) rather than filler. It is dense and uses long semicolon-joined sentences, which costs some readability, but 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?

An output schema exists, so return values need not be detailed, yet the description still previews the TRACES score, stuck points and replay link. Combined with the credit, prerequisite, and polling guidance, an agent has everything needed to invoke this correctly on the first attempt.

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 would be 3, but the description enriches the parameters beyond the schema: it explains the tier trade-off in cost terms, clarifies that email/password matter only for the signup template and continues into the signed-in home, and notes the purchase template needs no account. It slightly duplicates the schema text, which keeps it out of 5 territory.

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 ('Start a Menso test on a live website') and then explains the mechanism precisely: an AI persona in a real cloud browser attempting a task, scored 0-100 with a step replay. It also names the sibling tools in the workflow (poll get_status, then call get_findings), so an agent can place it relative to the rest of the toolset without opening other schemas.

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 selection criteria between the two templates ('purchase' needs no account; 'signup' requires email and password), states the URL must be publicly reachable and explicitly warns against localhost, and prescribes the follow-up pattern (poll get_status every 30-60 seconds, then get_findings). This is when-to-use, prerequisites, and sequencing all in one.

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