crm_set_classification

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

Annotations already declare write (readOnlyHint=false), non-idempotent, non-destructive and open-world. The description adds real value by disclosing the write target is channel memory rather than the authoritative pipeline, plus an implied revision-staleness failure mode. It stops short of saying what happens on a revision mismatch, whether prior labels are replaced, or whether the call is retryable.

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?

One sentence, front-loaded with what is stored, scope limit last. Dense noun stacking ('classifier provenance') costs a little readability but every clause carries information.

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?

With no output schema, three required parameters, a nested object, and zero schema descriptions, the description should do more: it omits the return value, revision-mismatch behavior, and whether existing classification is replaced. It credibly covers what is written and what is deliberately not touched.

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 0%, so the description carries the burden, and it partially does: 'expected thread revision' gives revision (an integer with only minimum 0 in the schema) concurrency semantics that the schema does not, and 'source-backed' hints at the evidence field. threadId and the nested rationale/classifierVersion/confidence fields remain unexplained, and the overwrite behavior of classification is unstated.

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 (store) and specific resource (source-backed labels, confidence, classifier provenance in LinkedIn channel memory). The trailing clause 'does not update the authoritative company pipeline' draws a sharp boundary against the CRM write siblings, though no sibling is named directly.

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 phrase 'at an expected thread revision' implies the caller must hold a current revision and that this is an optimistic-concurrency-aware write, and the 'does not update' clause implies when the tool is the wrong choice. But there is no explicit when-to-use statement, no pointer to crm_get_classification for reads, and no prerequisite/error guidance.

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