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

Annotations already declare non-destructive, non-readOnly, non-idempotent. The description goes beyond them with substantive policy: departures set status=departed and never delete the row (CRM-6), and provenance is correctable but never removable (CRM-9) — genuinely useful for an agent deciding whether a call is safe. It does not state whether omitted fields are left untouched, which matters given idempotentHint=false.

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 purpose and effect, and every sentence carries a rule or a constraint; nothing is padding. The rule codes (CRM-6, CRM-9) are cited without definition and the telegraphic phrasing ('confidence firms up') costs a little readability.

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 'Effects: contact updated' adequately covers the return. However, for a 13-parameter mutation tool the description never says whether the update is partial (omitted fields preserved) or a full replacement — a material omission given idempotentHint=false and the 'Correct or extend' framing.

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 77%, and the description still adds meaning for the highest-risk parameters: it explains the departed semantics, the correctable-but-not-removable rule for source, and reinforces the HUMAN-ONLY constraint on the three relationship-graph fields. Notes, title and confidence_note remain undocumented in both places, but the risky parameters are covered.

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?

Opens with a specific verb+resource ('Correct or extend a contact') and adds a concrete effect line ('Effects: contact updated'), which clearly separates it from crm_add_contact and the read-only crm_get_contact. It does not name those siblings explicitly, so the differentiation is inferable rather than stated.

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 sets a clear boundary between what the agent may change ('ordinary fields are open: titles change, confidence firms up, people depart') and what it must not touch (mutual_via, mutual_url, linkedin_path), with an explicit instruction to ask the human. It never mentions the sibling crm_set_path_in, which plausibly covers some of these relationship fields, so alternative routing is left unresolved.

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