submit_review

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

With no annotations, the description carries the full behavioral burden. It does disclose two valuable traits: originals are left immutable (non-destructive) and uncertain text stays in a pending state rather than being resolved. However, it says nothing about idempotency, whether a repeat submission overwrites a prior review, auth requirements, or failure behavior, which are significant gaps for a write tool with zero annotation coverage.

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 purpose and the key constraint are front-loaded in the first clause, and the field enumeration, while long, packs real structural information into a single compact list. The trailing sentence about immutability and pending text earns its place. Minor cost: the opening sentence is dense enough to require careful parsing.

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?

There is no output schema and no annotations, and the review object is a free-form nested structure, so the description is doing a lot of work. It documents the payload shape well and states the immutability/pending semantics, but it never explains what check_id is, what the tool returns, or what happens on repeat or invalid submissions, leaving the agent unable to fully predict the call's outcome.

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 0%, so the description must compensate, and it largely does by enumerating the nested review payload field-by-field (source_sha256, edits with index/before/after/reason/evidence/evidence_level/confidence, pending_review, overview with sections and limitations). The notable omission is check_id, a required parameter that is never mentioned or explained, so the compensation is substantial but incomplete.

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?

States a specific verb and resource ('Save a semantic review') and scopes it to content produced by the current host AI, so an agent immediately knows it is the persistence step for its own analysis rather than a fetch or query. It is unambiguous against siblings like get_report or get_transcript, though it never explicitly names an alternative, which is what separates a 4 from a 5.

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?

Provides a clear negative directive ('never call another LLM'), which tells the agent this is the terminal write step in the review workflow and not a delegation point. It gives solid context for when to call it, but names no sibling alternative or preconditions (e.g. requiring a prior check_id from check_creator), so it stops short of full when/when-not coverage.

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