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

Annotations already declare destructiveHint=true and idempotentHint=true, but the description adds behavior the annotations cannot: the memory "lasts across steps, compaction, turns, delegated agents and recovery," writes are revision-guarded ("a changed memory returns the current entries"), and unverifiable cites are silently dropped ("citesNotKept"). That discloses the non-obvious failure and persistence semantics an agent needs before mutating shared state.

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?

The opening clause is well front-loaded, but the rest is a single dense run-on sentence packing persistence, operation guidance, citation policy, concurrency, and prohibitions together. Nearly every sentence earns its content, yet the lack of separation into distinct lines makes the rules hard to scan and easy to miss.

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 complex, destructive, revision-guarded memory tool with an output schema, the description covers the write model, persistence scope, cite validation, and reply semantics — everything needed to invoke it correctly. Gaps are minor: request_key's idempotency role and read_skills' proof function are left entirely to the schema.

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 50% schema coverage, the description carries real weight and does: it explains expected_revision's role, the cite acceptance rules (workingMemory.commandId or other observed source for requirements/constraints; fullResultResourceId or resource@revision otherwise), and when final_reply must or must not be supplied. It does not address work_id, request_key, or read_skills beyond what the schema already says, so not a 5.

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 concrete verb-and-resource framing ("Keep this goal's notebook current") and enumerates the exact operations available — "Add what you found and decided; update an entry's status as it finishes; drop what no longer applies" — which matches the add/update/drop op enum. The metaphor "notebook" instead of "memory" and the absence of an explicit statement of what resource is being mutated keep it just short of a 5, but an agent can still tell what the tool does.

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 states a hard usage rule up front — call it "beside other calls or your final reply (final_reply), never alone" — plus per-operation guidance (add findings, update status when finished, drop stale entries). It also states exclusions: never record "hidden reasoning, credentials or user-facing replies," and it explains the revision precondition ("Pass expected_revision from the memory you read"). That is explicit when, how, and when-not guidance.

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