FlowAI Designer — Flow Design

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

Beyond the annotations, the description discloses genuinely important behavior: it is stateful and multi-turn ('il pose une question à la fois', results 's'affinent à chaque tour'), and the conversation is persisted by Flow Design for later resumption in the online atelier. This is consistent with readOnlyHint=false and idempotentHint=false and tells the agent why repeated calls with the same input are not equivalent.

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 what the tool is and what it returns, then the calling protocol, then persistence. Dense but every sentence carries information. Slightly heavy clause packing in the first sentence keeps it from a 5.

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 correctly compensates by describing the returned cadrage and déroulé horodaté. Turn mechanics and state persistence are covered. Minor ambiguity remains (whether the timestamped déroulé appears on every turn or only later, and any limits on turn count), but nothing an agent needs to call it 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 description coverage is already 100%, so the baseline is 3. The description still adds nuance beyond the schema by insisting the user's words be passed verbatim into 'message' and by tying 'conversation' to the post-first-call turn, clarifying sequencing that the schema states only structurally.

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 mechanism: a conversation turn with 'FlowAI Designer', the design agent, that returns a cadrage (objectifs, participants, format, livrables) plus a first timestamped déroulé. That is far more concrete than the name alone. It does not, however, explicitly contrast itself with siblings like parler_a_l_equipe or voir_seminaire, so an agent must infer that this one is the seminar-design conversational entry point.

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 operational context: pass the user's own words in 'message', and after the first call pass the 'conversation' identifier. That effectively defines when to omit the parameter (start a new seminar). It stops short of explicit when-to-use/when-not guidance relative to the two sibling tools, which would have earned 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 goes well beyond the annotations: it discloses third-party data transmission to an external team, the consent gate, the no-quote policy, and a rate limit ('au plus toutes les 10 minutes') that explains the non-idempotent hint. What it omits is the caller's side of the interaction — what the tool returns or confirms after the handoff.

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?

It leads with what the tool does before layering on the consent rule, the exclusion and the rate limit, which is the right order for an agent deciding whether to call. It is a dense single paragraph, and the '« consentement » : true' fragment mildly duplicates the schema's const constraint.

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 9-parameter, 6-required mutation tool with no output schema and a light annotation set, the description covers the critical agent decisions: consent, data provenance, external transmission, and re-call behavior. The remaining gaps — the shape of the confirmation returned, and what the optional langue and message fields are used for — are 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?

With 67% schema coverage, the schema already documents email, message, telephone, entreprise and conversation. The description nonetheless adds real meaning: it defines the consent semantics more strictly than the schema does (explicit agreement given in the conversation, for those specific pieces of information) and clarifies that 'le cadrage et le déroulé' travel with the request, i.e. what the conversation identifier carries.

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 precise action (putting the user in contact with the Flow Design team), names the recipients (Victoire and Marie-Violaine), and specifies exactly what they receive (contact details, framing, conversation flow) and how they respond (by e-mail). It is unmistakably distinct from the seminar-design siblings, which deal with entirely different resources.

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 an explicit precondition ('À n'appeler qu'avec l'accord explicite de l'utilisateur... consentement: true'), an explicit data-source rule (only contact details the user has just provided), and an explicit exclusion ('Pas de devis : l'équipe chiffre après l'échange'). Repetition behavior is also flagged, so the agent knows calling twice has effects.

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, idempotentHint=true, destructiveHint=false and openWorldHint=false, so the safety profile is fully covered. The description adds real value by disclosing the shape of the return: current framing and flow plus a link to resume the conversation in the Flow Design workshop. It says nothing about error behavior when a conversation ID is invalid or unknown.

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 compact sentence with no filler, and the most informative element (the resume link) is placed at the end. Its weakness is structural rather than verbosity: it lacks a leading verb to anchor what the tool does.

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 two-parameter read-only tool with annotations covering safety, the description is nearly sufficient, and it usefully compensates for the absence of an output schema by naming the returned fields. It falls short on parameter meaning, leaving both conversation and langue undocumented alongside a 0% schema coverage.

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 full burden. It alludes to "une conversation avec FlowAI Designer", which loosely maps to the required conversation parameter, but never explains that the parameter is a 22-character conversation identifier. The langue parameter (fr/en enum) is entirely unmentioned, as is its default or effect.

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 is a noun phrase describing the content returned ("le cadrage et le déroulé actuels d'une conversation... et le lien pour la reprendre") rather than naming an action verb. It does go beyond the bare name voir_seminaire by specifying that the output is the conversation's framing, flow, and a resume link, but an agent must infer that the tool retrieves this data. It does not differentiate itself from the sibling concevoir_seminaire.

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 explicit statement of when to use this tool versus concevoir_seminaire or parler_a_l_equipe. The reader can infer it is for reviewing an existing seminar conversation, but no conditions, prerequisites, or exclusions are stated. The only guidance is implicit in the phrase "reprendre dans l'atelier Flow Design".

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