sensitivity

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

Annotations declare readOnlyHint=true, so the agent knows this is a safe read operation. The description adds value by explaining what the tool computes (bounds on unobserved confounding strength) and that it supports multiple methods. It also mentions the 'detail' parameter's 'agent' mode adds violations/next_steps/suggested_functions, which is behavioral context beyond the schema. 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 two sentences, front-loaded with the tool's purpose and methods. The second sentence adds the key interpretive value ('how strong an unobserved confounder would have to be to overturn the result'). No wasted words, though it could be slightly more structured with explicit method names in a list.

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?

The tool has 14 parameters, an output schema, and annotations. The description explains the core purpose and the 'detail' parameter's role in agent planning. It doesn't explain the difference between 'result' and 'result_id' or when to use data_path vs result_id, but the schema covers those. The output schema exists, so return values need not be described. Complete enough for an agent to select and invoke correctly.

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 79%, so the schema already documents most parameters. The description adds context for the 'result' parameter (bounds are computed relative to this estimate) and the 'detail' parameter (payload depth for LLM planning). However, it doesn't explain the relationship between 'result' and 'result_id' or clarify which parameters are needed for a first call vs a chained call. Baseline 3 is appropriate given high schema coverage.

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 ('supports') and resource ('sensitivity analysis for observational causal estimates'), names three concrete methods (Oster, Cinelli-Hazlett, E-values), and states the tool's purpose: telling the agent how strong an unobserved confounder would have to be to overturn the result. This clearly distinguishes it from siblings like sensitivity_dashboard, sensitivity_plot, and oster_bounds.

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 implies when to use this tool: when the agent needs to assess robustness of an observational causal estimate to unobserved confounding. It names the methods it supports, which helps an agent choose it over alternatives. However, it does not explicitly state when NOT to use it or name alternative tools (e.g., sensitivity_dashboard, sensitivity_plot, oster_bounds) for different needs.

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