The Claude model for a helper

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

Annotations declare readOnlyHint=false, so the agent knows this is not a pure read, and the description hints that done/turn payloads record usage — roughly consistent, but it never says what a write does, whether it is persisted, or what comes back. 'Numbers only, never words' is cryptic and could describe the return value or the inputs; either way it is not actionable. With no output schema and a mutation-flagged tool, this leaves a substantial 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?

The text is a dense run-on with an unclear antecedent structure and jargon ('leanfare-level mod', 'the plan part it is for or else its kind of work'), forcing re-reads. The key purpose is buried mid-sentence rather than front-loaded, and several clauses do not clearly earn their place.

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 7-parameter tool with nested objects, no output schema, and a mutation flag, the description should at minimum explain the pick-vs-record modes and the returned model-name shape. Instead it offers a cryptic 'Numbers only, never words' and no statement of persistence, idempotency behavior, or error cases, leaving the agent unable to call it confidently.

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 per-field meanings of done, task, turn, agent, asked, session and description are already documented. The description adds only marginal context beyond that, chiefly the selection rule ('by the plan part it is for or else its kind of work'), which is useful but underspecified about precedence. Baseline 3 is appropriate when the schema carries the parameter load.

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 does eventually convey the core purpose: return the Claude model Leanfare would pick for a subagent, named as the Agent tool names them (haiku, sonnet, opus, fable). But the opening clause ('For Leanfare's leanfare-level mod and Claude Code's hook, not for a chat') is internal jargon that obscures the verb, and mixing selection with recording ('with done... with turn...') makes it ambiguous whether this is a lookup or a telemetry sink. The sibling tools (models, delegate, plan) are not referenced, so differentiation is left to inference.

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?

The only usage signal is the exclusion 'not for a chat', with no positive statement of when to call this instead of sibling tools like models or delegate. No prerequisites, ordering, or call-frequency guidance is given, and the two modes (pick a model vs. report done/turn usage) are not tied to any condition.

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