talon_scan

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

Annotations already declare readOnly/openWorld/idempotent, so the bar is lowered, and this description still adds substantial context: findings are never dropped, taste-binding overrides return findings flagged as waived_by_taste:true, geometry rules require premeasured elements+viewport, and every finding is traceable to a principles source.

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 information-dense, front-loaded, well worth its length, and reads naturally. It earns a 4 rather than 5 because the dense enumeration of rule categories and nested parentheticals makes the paragraph quite long and less skimmable than a more structured summary.

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?

Even with no output schema, the description effectively conveys what a caller should expect: findings with citations, possible waived flags, and no dropping. It also covers the interaction between all meaningful parameters and the conditions under which features activate, making it complete enough to select and invoke safely.

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 baseline is 3, but the description goes far beyond the schema by explaining which input mode activates which rules, how project/profile/waivers affect results, and that surface scoping is optional. This is the kind of semantic gloss that makes the tool safe and effective.

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+resource: runs Raven's deterministic detector engine over a page, and clearly distinguishes itself from LLM-based evaluation by saying 'no LLM, pure measurement.' It lists the concrete rule domains, so an agent knows exactly what class of checks this tool performs.

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?

Provides clear context: pure/deterministic measurement, three input modalities, conditions for geometry rules, project profile resolution, and surface scoping. What keeps it from a 5 is the absence of explicit alternative-routing ('use audit_page for X instead' or a named sibling comparison).

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