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

Annotations already declare readOnly/non-destructive/closed-world, so the bar is low, yet the description adds substantial behavioral detail: the tool is metadata-only and does not fetch real rows, signals carry trust caveats (weak joins, fields not to quote back), statements_only returns {} when nothing applies, and the cursor exists because the client may cut a long answer. That is real context an agent cannot get from 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?

It is front-loaded with the key instruction and roughly the right length, but the middle is a long semicolon-separated dump that mixes mode contents, caveats and warnings into one undifferentiated block. Splitting the three modes into distinct sentences would cost nothing and greatly improve scanability.

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 nine-parameter tool with nested objects, three mutually exclusive modes and no output schema, the description covers the critical ground: mode selection, the metadata-only boundary, the safety framing around suggested_action, truncation/pagination behavior, and a doc pointer for depth. The main gap is that it never sketches the shape of the returned context, which the absent output schema would otherwise require.

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 3 is the baseline, but the description goes further by explaining what each mode yields (task → objects/routes/playbooks/owners; signal_id → evidence/remedy/handling/trust/answer_with/approved_fix) and adds ranking-vs-filtering semantics for statements_query plus the 'keep the same query' constraint on statements_cursor. Only rule_node, statement_key and context are left largely to the schema.

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 gives a specific verb and scope ('one call equips you for a task, signal or support case') and then enumerates exactly what each of the three input modes returns. It is clearly a context-priming tool distinct from list_signals or get_object, though the dense semicolon chains and internal jargon (layer, rests_on, coverage, weak joins) make the purpose slightly harder to absorb than it needs to be.

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?

It states plainly when to call it ('Start any work here'), routes the agent elsewhere where appropriate ('fields via get_object', 'open a rule's source only if unclear'), and sets ordering rules ('read payload.checks before acting on one'). It also draws a hard boundary against misuse ('Still metadata: fetch rows in your human's own systems'; 'never a permission'), which is exactly the guidance an agent needs before acting.

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