Read messages

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

Annotations already include readOnlyHint=true (the tool is read-only) and openWorldHint=true (results may change between calls). The description goes beyond this by disclosing that reading doesn't consume send quota, that messages have a few seconds of delay, and that 'from' is inclusive while 'to' is exclusive to avoid double-counting. It also mentions that encrypted messages cannot be searched. These behavioral traits are not implied by the annotations and add significant value.

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 detailed but well-organized, opening with the core purpose and then layering usage patterns (defaults, polling, batch checks). It is front-loaded with essential info (no device id needed, direction default) and later details are relevant to specific use cases. It is longer than ideal but every sentence adds value, and the structure helps the agent parse the key information without unnecessary fluff.

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 the tool's complexity (9 parameters, no output schema, no required params), the description is comprehensive: it covers direction semantics, polling mechanics, batch handling, device filtering, and search limitations. The absence of an output schema is partly mitigated by describing what results include (delivery status per row, next_cursor). Minor gaps include not specifying pagination details for direction='all' or explicitly saying that results are ordered, but the overall completeness is strong.

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%, so the schema already documents all nine parameters. The description adds crucial semantics beyond the schema: it clarifies the default direction and its rationale, explains that 'from' is inclusive and 'to' is exclusive for boundary handling, and describes the cursor's role in polling. It also integrates parameters like sms_batch_id and device_ids into workflows. The description substantially enhances the agent's understanding, justifying a score above the 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 clearly states the tool reads SMS messages from the user's textbee account, covering all devices without needing device IDs. It explicitly contrasts with siblings: it is for reading, not sending (send_sms) or listing devices (list_devices). The scope, direction defaults, and batch filtering are all specified, making its purpose unambiguous and distinct.

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 provides rich usage guidance: it says when to use the default direction 'received' (for replies or one-time codes), when to use 'sent' (reviewing outgoing), and when to use 'all' or 'sms_batch_id' (checking delivery). It describes a complete polling pattern with order, from, and cursor, and explicitly advises waiting and retrying instead of tight-polling. No siblings are alternatives for reading, but the description covers all relevant scenarios.

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