spillover

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

Annotations only declare readOnlyHint:true and openWorldHint:false. The description adds substantial behavioral context beyond these: assumptions (partial interference, correct exposure function, overlap), failure modes (misspecified exposure_fn), validation tier, and typical minimum N. It does not contradict annotations; the read-only hint aligns with estimation-only behavior. The extra context fully discloses edge cases and limitations.

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 organized into labeled sections (assumptions, pre-conditions, failure modes, alternatives) and front-loads the core purpose. Every sentence carries information, but it is somewhat dense and could be seen as longer than necessary. Still, the structure improves scannability and no information is redundant. It is concise relative to the breadth of content.

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?

For a complex estimator with 13 parameters and 4 required, the description covers purpose, assumptions, pre-conditions, failure modes, alternatives, and typical sample size. It leverages the output schema to avoid explaining return values. No critical usage aspect is missing; an agent could correctly invoke it for partial interference scenarios and know when to use alternatives.

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 baseline is 3. The description adds limited parameter-specific meaning: it references exposure_fn (fraction/any/count) in assumptions and failure modes, and clarifies the cluster parameter as defining the interference boundary. However, it does not elaborate on other parameters (y, treat, data_path, etc.) beyond what the schema already states. It adds marginal value, not enough to push beyond 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?

States a specific verb and resource: 'Direct + spillover treatment effect estimation under partial interference (within-cluster). Uses the Hudgens-Halloran decomposition with chosen exposure function.' This clearly distinguishes it from other causal tools by naming the method and interference assumption. It also explicitly lists alternatives (sp.network_exposure, sp.cluster_matched_pair, sp.peer_effects), making differentiation unambiguous.

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 explicit when-to-use conditions: pre-conditions (cluster column, within-cluster variation, >=30 clusters) and failure modes that route to alternatives (e.g., 'No within-cluster variation in treatment -> use sp.cluster_matched_pair or cluster-level ATE'). Also names alternatives directly, leaving nothing to inference.

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