Customer Intelligence & Segmentation

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

Annotations already declare readOnlyHint=true and destructiveHint=false, and the description adds substantial behavioral context beyond that: ID variants are merged server-side by rule, ambiguous identities are surfaced unmerged in review_candidates, rep_contacts learning is 'observational, not causal', and full result includes per-customer ledgers with explanation traces. 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 long but every paragraph earns its place: core action, size limit and sibling routing, optional rep_contacts, and the identity-cleaning workflow are all present. Some instructions repeat the schema (e.g., 'do not pre-normalise'), but the repetition is intentional emphasis rather than filler.

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 tool with 16 parameters, nested objects, and an output schema, the description is unusually complete. It covers the pipeline, the boundary condition for the sibling tool, the identity-cleaning model the agent must understand, and where ambiguous cases surface. The output schema covers return values, so further return-format detail is unnecessary.

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 is 3. The description adds meaningful semantics for key parameters: rep_contacts learns uplift from data rather than assuming it, transaction customer_id must be sent exactly as sourced without pre-normalisation, customer_name is only for reporting labels, and identity_overrides is for judgement calls on review_candidates. This elevates it above baseline.

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 opens with a specific verb and resource: 'Scores a customer book of 500 transaction rows OR FEWER into A/B/C/D tiers.' It names the exact models run and what is returned (per-customer ledger with explanation traces), and it clearly distinguishes itself from the sibling by the 500-row boundary.

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?

Explicit when-to-use guidance is present: 'For books LARGER than 500 rows use customer_tiering_get_engine instead — sending thousands of rows as tool arguments is slow and risks truncated JSON.' It also tells agents not to pre-normalise identifiers and to use identity_overrides only after judging review_candidates, which prevents misuse.

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

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

Annotations already mark the tool readOnly and non-destructive, but the description adds substantial behavioral context: rows never pass through as tokens, no customer data reaches the server, the script writes a result JSON, requires numpy, cleans identities itself, and prints merged variants for review. This is far beyond what annotations provide.

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 long but every sentence carries operational weight. It front-loads the threshold, the deliverable, and the key warning, then systematically covers constraints, optional inputs, and output behavior. The repeated emphasis on running the script verbatim is justified given the likely failure mode.

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 tool that returns a script, the description covers prerequisites, execution environment, required input format, input-size rationale, optional input value, output behavior, and anti-patterns. The presence of an output schema means return values do not need to be inventoried again. Nothing essential is missing.

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 is 3. The description adds meaning beyond the schema for file_path (do not pre-clean, purchase timing required) and especially contacts_path (explains why it is valuable and when to ask for it). as_of and currency are only covered by the schema, so the description does not fully elevate to 5.

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 names a specific resource ('complete, runnable Python script'), a precise action ('scores the book into A/B/C/D tiers'), and a clear scope condition ('books LARGER than 500 transaction rows'). It also enumerates the methods, making the tool's purpose impossible to confuse with a generic scoring tool.

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 gives explicit when-to-use guidance: books larger than 500 transaction rows. It also warns against pre-cleaning the file, instructs the agent to run the returned script verbatim, and specifies when to ask for contacts_path ('whenever the user mentions a CRM, a call log or a visit register'). This is strong, actionable routing.

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