trace_failed_tools

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

With no annotations provided, the description carries the full burden of behavioral disclosure, and it discloses almost nothing beyond the core read operation. It does not define what 'failed' means (errors, exceptions, timeouts, cancellations), whether 'recent' implies a bounded time window, what ordering is used, or whether there is pagination. No contradictions exist, but the behavioral surface is essentially bare.

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 a single six-word sentence with zero filler, and the verb 'Return' is front-loaded. It is efficient for what it says, though brevity here edges toward under-specification rather than disciplined conciseness.

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?

An output schema exists, so return structure is covered elsewhere, and the tool has only one optional parameter. But the two key semantic terms, 'recent' and 'failed', are left undefined: the agent cannot know the recency window or what failure criterion determines inclusion, which materially affects correct interpretation of results.

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 should compensate but does not mention the single `limit` parameter at all. However, the parameter is low-risk: its name 'limit' and integer type with default 20 make the meaning (cap the number of returned spans) nearly self-evident, so an agent is unlikely to misuse it.

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 states a specific verb ('Return'), a specific resource ('recent spans'), and a filter ('that failed'). This distinguishes it from siblings like trace_token_cost and trace_slowest_tools, which target cost and latency respectively. However, it does not fully differentiate from all siblings, since 'failed' is not defined and trace_list_runs or trace_get_run could overlap in some interpretations.

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?

There is no guidance on when to use this tool versus the six siblings. No alternatives are named, no conditions are given (e.g., 'use trace_slowest_tools for latency debugging'), and an agent debugging a run has no stated basis for choosing 'failed spans' over 'list runs' or 'get run'. All usage context must be inferred from the tool name alone.

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