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

Annotations already declare idempotentHint=true and destructiveHint=false, so the retry sentence is largely reinforcement, but it adds specificity about what retrying returns. The genuinely new disclosures are the 20-memory cap and the dedupe behavior, which are not in the structured fields. It stops short of saying what happens when the cap is reached (error vs. eviction), which is the main remaining behavioral gap.

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 with no filler; the core action and the topic taxonomy come first, followed by the cap and retry semantics. Slightly packed into one flowing sentence, but every clause carries information.

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 create tool with no output schema, the description covers the essentials but leaves real gaps: no indication of what is returned (the created memory, an id, or nothing) and no statement of what happens at the 20-item cap. An agent could call it correctly but might be surprised at the boundary.

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 50%, with the nested body object undocumented, so the description has to compensate – and it does partly, glossing 'content' as 'one line in your own words' and clarifying that 'icp' means who to target. The idempotencyKey already carries its own schema description, so that adds nothing beyond the schema; the 300-char limit and exact topic semantics are left to the 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 gives a specific verb and resource ('Saves one lasting preference') and scopes it to the workspace, plus names the three topic buckets (messaging, icp, other). It implicitly separates itself from updateAgentMemory and listAgentMemories but never names those siblings, so an agent must infer the create-vs-update boundary.

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 implied rather than stated: the word 'Saves one lasting preference' suggests creating a new memory, and 'Up to 20' hints at a limit, but there is no explicit 'use this when X, use updateAgentMemory when Y' routing. The dedupe note ('saving the same preference twice keeps one') does give useful behavioral conditions for calling it.

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