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

Annotations provide readOnlyHint=true and destructiveHint=false, but the description adds substantial behavioral context beyond them: the signed page_token is 'source-pinned', a controlled in-period centre correction 'changes a displayed label and invalidates the page token' while 'persisted allocation amounts stay unchanged', staleness/methodology changes require regeneration, integrity failure prescribes regenerate → retry → contact support, and the return excludes 'owners, descriptions, source transactions, or recomputation.' It also clarifies numeric semantics ('Accounted equals assigned plus Unallocated; Unallocated is not additional'). This is far beyond what annotations alone convey.

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?

At roughly 250 words, it is long, but nearly every sentence carries operational weight — paging, reconciliation non-additivity, step-down semantics, label timelines, failure handling, and return exclusions. The purpose is front-loaded in the first sentence. It is not optimally organized (a stream of caveats rather than clearly grouped sections), and a few edge-case details (target_labels_frozen, controlled in-period corrections) could be trimmed, but nothing is redundant fluff.

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 takes on the burden of describing returns and does so well: it states the grain, the decimal-string amount format, and explicitly what is not returned ('No owners, descriptions, source transactions, or recomputation are returned'). It covers pagination statefulness, failure behavior, and data semantics. Minor residual gaps: no explicit JSON response shape, no error-code enumeration, and the 'mode' component of the grain is never defined. Very strong coverage for a complex 9-parameter finance tool, but not exhaustive.

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 100%, so the baseline is 3; the description adds genuine value on top: 'For profit centres, choose profit' reinforces allocation_dimension; 'do not add them [the two views]' gives semantic context for cost vs profit; 'Filter allocation_sources=['step_down'] only to inspect transfers and confirm their run net is zero' gives purpose-specific filter semantics; and 'restart at page 1 if source_changed' adds behavioral meaning to page_token. This exceeds the baseline but the schema already documents each parameter well, so it does not reach 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?

Opens with 'MANAGEMENT data class. Read one explicit cost or profit partition of the current persisted allocation run at account × target × source × mode grain, with decimal-string allocated amounts.' This is a specific verb+resource statement that names the grain, the persistence scope, and the amount format — far beyond a generic 'reads allocation figures.' The phrase 'cost or profit partition' and 'target figures' distinguish it from the sibling get_allocation_account_figures, even without naming it.

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?

Rich when-to-use guidance: 'Filter allocation_sources=['step_down'] only to inspect transfers and confirm their run net is zero', 'For profit centres, choose profit', 'Use filters before paging', 'do not add them' (the two reconciliations), and restart-at-page-1 rules for source changes and centre corrections. The only gap is that no sibling alternative is named explicitly — e.g., it never says 'for account-level figures use get_allocation_account_figures' — so it stops short of a 5.

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