Remember a durable fact

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

Annotations already declare write (readOnlyHint=false), idempotent, non-destructive, and the description usefully adds the mechanism behind idempotency: "Identical entries are deduplicated." It also discloses that entries land in a local project store unless scope is global. It does not say what a caller gets back or how deduplication resolves (silent skip vs. error), which keeps it off a 5.

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, both load-bearing: the first establishes what is written and where, the second establishes when to write. The dedup note is a single trailing clause rather than its own paragraph. Nothing is repeated or padded.

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 write tool with no output schema, the description covers purpose, storage scope, triggers, and dedup behavior, which is most of what a caller needs. It omits how a caller obtains the ID required by `replaces`, and it leaves the relationship to `recall`/`forget` implicit, which is a genuine gap for a memory system with sibling tools.

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 78%, so the schema already documents most of the 9 parameters, and the description's enumeration of types and the project/global distinction largely restates the enum values. It adds no extra semantics for tags, files, importance, source, or replaces beyond what the schema descriptions provide. Baseline 3 is the right ceiling here.

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 concrete verb (persist) and an enumerated resource set (decision, gotcha, fact, command, todo, note) plus the store it lands in, so an agent immediately knows what the tool does. It falls short of a 5 because it does not distinguish itself from the `todo` sibling even though it accepts a "todo" type, leaving a real ambiguity unaddressed.

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?

"Use this whenever the user states a preference, you make a non-obvious choice, or you discover something future sessions should know" gives three explicit, checkable triggers. It stops short of the top score because it never names alternatives or exclusions — nothing says to use `recall` to look things up, `forget` to remove, or `todo` for tracked tasks.

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