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

Annotations declare destructiveHint=true, openWorldHint=true, and readOnlyHint=false, which are significant traits the description never echoes or explains — an agent gets no warning about what is destroyed or that processing goes to an external inference provider, despite consentExternalProcessing being a required param. The description does not contradict the annotations, but it adds almost no behavioral context beyond restating 'from your supplied sources'.

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 no filler and the core operation front-loaded. Brevity is appropriate in form, though at this length it shades into under-specification for a six-parameter, nested, open-world tool.

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 six parameters, a nested inputs object with a 75,000-character combined limit, no output schema, destructive and open-world annotations, and required consent plus idempotency key, the description is far too thin. It does not explain what comes back, how sourceRunIds reuse interacts with fresh inputs, or what the caller commits to by invoking it.

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 67%, below the threshold where the schema carries the load, and the description compensates for none of it. It never explains the relationship between `inputs` (raw evidence), `sourceFields`/`sourceRunIds` (reuse of prior runs), or why `consentExternalProcessing` must be literally true — the single most consequential parameter for the caller.

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 verb 'synthesize' plus the resources (procurement, market, service knowledge) give a rough sense of intent, and the optional provenance-tagged knowledge graph is a concrete artifact. But the actual deliverable is never stated (a brief? a set of findings? a graph?), and nothing distinguishes this from the many category-pack siblings (mro, indirect, it-software) or from procurement_workflow__knowledge_to_negotiation. Purpose is implied rather than pinned down.

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 when-to-use guidance, no when-not-to-use, and no named alternative. With ~30 siblings including category-specific tools and a workflow tool literally named knowledge_to_negotiation, the absence of routing information is a real gap. The two sentences describe what happens, not when to pick this tool.

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