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

Annotations declare destructiveHint=true, and the description elaborates: 'Execute spends operator-authorized credits; resume is free.' It also clarifies reuse ('Reuses unchanged partitions') and constraints ('No arbitrary code or semantic inference'). This adds meaningful behavioral context beyond the annotations, such as the credit cost and the requirement for expected_version on writes. No contradiction with annotations.

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 description is a single, dense paragraph that front-loads the primary actions and then adds key constraints. It is concise and avoids fluff. The structure flows logically from actions to behavioral notes. It could be slightly more organized (e.g., bullet points), but it is efficient and readable.

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?

Given the tool's complexity (7 parameters, nested objects, multiple actions, no output schema), the description is incomplete. It does not explain what the response looks like for each action, nor does it cover error handling or edge cases. It mentions match returns 'a compact existing result' but omits details for quote, execute, and resume. The description also fails to clarify the distinction between 'plan' and 'match_plan' or the role of 'max_credits' and 'idempotency_key'. This is a significant gap for an agent to call the tool correctly.

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 for undocumented parameters. It mentions 'expected_version' and 'source guards' but does not explain the meaning of 'contract', 'sources', 'snapshot_complete', 'reuse_policy', or other fields. The description's mention of 'exact source scope' hints at the sources parameter but lacks specificity. It fails to add semantic value for most parameters, leaving the agent to infer from the schema alone.

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 clearly states the tool's function: 'Quote, execute, resume or match a deterministic table capsule.' It enumerates four distinct actions and specifies that match 'validates an exact source scope and returns a compact existing result without debit.' This distinguishes it from sibling tools like mithrandir_query and mithrandir_read, which likely handle general queries. The phrase 'No arbitrary code or semantic inference' further differentiates it as a deterministic operation.

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 description provides some usage context: match returns existing results without debit, execute spends credits, resume is free, and writes require 'explicit complete scope, source guards and expected_version.' However, it does not explicitly state when to prefer this tool over siblings like mithrandir_query or mithrandir_read, nor does it mention exclusions. The guidance is mostly internal to the four actions, not a comparison with alternatives.

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