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

Annotations already declare readOnlyHint=false, destructiveHint=false and openWorldHint=false, so the safety profile is covered. The description adds persistence semantics — the saved request becomes version one and enables rerun/share/change — which is genuinely useful context beyond the annotations. It omits idempotency, naming/duplicate behavior, permission requirements, and error modes for a mutation tool.

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?

A single, tightly written sentence with no filler, front-loaded on the action, and the trailing "so it can be rerun, shared or changed" earns its place by explaining the value of saving. It is efficient but borders on under-specified rather than merely concise.

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?

This is a complex tool — 6 parameters, a deeply nested QueryAst with dozens of $defs, a mutually exclusive ast/source_find_id requirement, and no output schema. One sentence cannot cover the input contract, and the description leaves the agent unprepared for how to build the payload or what the call returns.

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 only 17% across 6 parameters, and the description adds no parameter meaning at all — it never mentions name, q, ast, source_find_id, project_id, or context. Most critically, the schema's oneOf requiring exactly one of ast or source_find_id is not surfaced in prose, so the agent gets no help on the tool's central input decision.

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 a reviewed request") and adds the key differentiator "as its first version", which separates it from revise_spec/edit_spec. It stops short of naming any sibling outright, so an agent must infer the create-vs-revise 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?

"Reviewed request" implies a prerequisite (the request must already have been reviewed, presumably via compile) before saving, which is useful implicit guidance. However, there is no statement about when to use this versus revise_spec, edit_spec, archive_spec, or create_spec_link, and no note about what to do if a spec with the same name exists.

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