bcf_ordinal

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

Annotations declare readOnlyHint=true, so the agent knows this is a read-only estimation call. The description adds valuable behavioral context beyond annotations: it discloses that a propensity model is fit internally to limit regularization-induced confounding, that MCMC diagnostics may fail to converge, and that estimates may be prior-sensitive. It also describes the chained-BCF estimation strategy. It does not detail output structure, but the output schema exists and the detail parameter covers payload depth.

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 dense but well-organized with labeled sections (Assumptions, Pre-conditions, Failure modes, Alternatives, Typical minimum N). Every sentence carries information. It is somewhat long, but the complexity of a BCF tool with MCMC diagnostics justifies the length. The most important identifying information (ordered/dose-level, cumulative dose-response) is front-loaded.

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 (16 params, MCMC, BCF), the description covers the key contextual needs: what it estimates, assumptions, pre-conditions, failure modes, alternatives, and sample size guidance. The output schema and detail parameter handle return-value expectations. It could be more complete with explicit guidance on which parameters to tune when diagnostics fail, but the failure-mode section already points to draws/tuning and outcome rescaling.

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 16 parameters. The description adds context for the treatment variable (ordered/dose-level) and mentions internal propensity fitting, which helps interpret treat and covariates. However, it does not add meaning beyond the schema for most parameters (e.g., n_trees_mu, n_trees_tau, n_bootstrap, alpha, baseline). Baseline 3 is appropriate given full 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 clearly identifies the tool as a Bayesian Causal Forest for ordered/dose-level treatment, specifies the method (chained BCF between consecutive levels), and states the estimand (cumulative dose-response curves). It distinguishes itself from siblings like bcf, bcf_factor_exposure, bcf_longitudinal, and causal_forest by the ordered/dose-level focus.

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 explicit pre-conditions (covariates, treatment, outcome; propensity model fit internally), assumptions (unconfoundedness, overlap, BART priors appropriate), and failure-mode guidance (increase draws/tuning, re-scale outcome, report diagnostics). It names alternatives (sp.dml, sp.auto_cate, sp.causal_forest) and gives a minimum N of 250. It does not explicitly state when NOT to use this tool versus those alternatives, but the context is strong.

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