cannibalization_check

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

With no annotations provided, the description carries the full burden, and it largely delivers. The POSITION paragraph is a substantive behavioral disclosure: position is an impression-weighted average blended across queries/devices/countries, not a rank-tracker rank, deep positions rest on sparse data, and absolute values are directional. The IMPORTANT section constrains the agent to report only returned numbers without speculation. It stops short of disclosing return structure or data-source specifics, but the metric caveats are genuinely valuable beyond what schema or annotations could convey.

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 (~170 words) but well-structured: purpose is front-loaded, and the IMPORTANT/PRESENTATION/POSITION sections are clearly delineated with capitalized headers. Each section serves a distinct need. The PRESENTATION section is the most expansive — specifying cards, color-coding, sparklines, and tabs — and could be trimmed, but it earns its place for a tool whose output is meant to be dashboard-like.

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 there is no output schema, the description reasonably explains the key output concepts (winning page, position gap, impressions split) plus the crucial interpretation caveat about position's metric semantics. It covers purpose, data-fidelity rules, presentation format, and metric interpretation. The main gaps are the absence of a concrete return structure/field names (no output schema to lean on) and no parameter usage guidance, but for a 2-optional-param read tool, the description is largely sufficient.

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 100% (both 'days' and 'min_impressions' have descriptions), so the baseline is 3. The tool description adds no parameter-level meaning: it never explains tradeoffs, suggested ranges, or how min_impressions relates to the 'combined impressions being split' output concept. The schema's one-line descriptions carry the entire parameter burden.

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 — 'Find keywords where multiple pages from your site compete against each other' — uses a specific verb and resource, and the second sentence details the outputs (which page ranks higher, position gap, combined impressions split). This is inherently distinct from all sibling tools; nothing in the sibling set (content_gaps, traffic_drops, ctr_opportunities, etc.) addresses internal page competition for the same keyword.

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 no explicit tool-selection guidance: it never names an alternative tool, states when NOT to use it, or contrasts with siblings. The IMPORTANT, PRESENTATION, and POSITION sections are extensive but are post-invocation behavioral instructions (how to interpret and present results), not guidance on when to choose this tool. The use case is only implied by the purpose statement.

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