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

Annotations only declare it as a non-destructive write; the description adds substantial behavior the annotations do not: a named refusal if the task doesn't exist, the explicit no-cascade rule so threads survive task deletion, and how list_comments surfaces such threads as orphaned:true. It also discloses that mention IDs without real project access are silently dropped and reported in a `warning`, which is real anti-surprise information.

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?

Front-loaded with the core action, then a dense anchor-type enumeration where each clause carries distinct information. It is long and would scan better as a per-anchor list rather than one paragraph, but there is little redundancy to cut.

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 nested-anchor mutation tool with six params and no output schema, the definition covers preconditions, the anchor decision tree, and failure/orphan semantics. It says almost nothing about the successful return value (thread id, the mentioned `warning` field is only mentioned in passing), which is the one thing an agent would still need to guess.

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 coverage is 100%, so the baseline would be 3, but the description goes well beyond the schema's generic 'Comment anchor (...)' sentence: it maps each anchor type to its required companion IDs (pinX/pinY for free, documentId alone for document_range, taskId alone with explicitly no cellId for task) and explains why. The schema declares none of these conditional requirements.

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 opening sentence is a precise verb+resource: 'Creates a new comment thread + first message.' It also enumerates the anchor variants, which clarifies the tool's scope. However, it never distinguishes this tool from the sibling reply_comment, which is the obvious alternative when a thread already exists — the agent must infer that this one starts a thread rather than extending one.

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 bulk of the text is guidance on which anchor.type to choose for a given situation, which is genuinely useful routing. It also states a precondition ('The task must exist in the active project (named refusal otherwise)'). What is missing is the primary when-to-use decision: new thread here vs. reply_comment for an existing thread, and it never says what to do when you don't have a cellId.

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