Causal Decision Agent

Overview Schema Source Related Servers Discussions Admin

tmle

Read-only

Estimate causal effects (ATE/ATT) with doubly robust targeted estimation using binary treatment, outcome, and covariates from a dataset.

Instructions

Targeted Maximum Likelihood Estimation for ATE/ATT with double-robustness. Validation: certified parity evidence. Assumptions: Unconfoundedness: Y(d) perp D | X; Overlap: 0 < P(D=1 | X) < 1 on the estimand support; Consistent estimation of at least one of Q(a, x) = E[Y|A, X] or g(x) = P(A=1|X) (double robustness). Pre-conditions: binary treatment 0/1; covariates comprise the confounding set; n >= 500 for asymptotic efficiency. Failure modes: Extreme propensity scores (ATE IF denominator ~ 0) -> Bound propensity scores away from 0/1 (e.g. 0.025 / 0.975) or trim; Super-learner cross-validated risk not improving over baseline -> Nuisances not learnable; widen the candidate library or use stronger base learners. Alternatives: sp.dml, sp.aipw, sp.metalearner, sp.ltmle. Typical minimum N: 500.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`y`	Yes	Outcome variable column name or outcome array.
`treat`	Yes	Binary treatment column (0/1)
`detail`	No	Payload depth: 'minimal' (~150 tokens) for sub-step calls where only the point estimate is needed; 'standard' (~1K tokens) for diagnostics + coefficient table; 'agent' (~2K tokens, default) adds violations / next_steps / suggested_functions so the LLM can plan its next call without another round-trip.	agent
`estimand`	No	Target estimand	ATE
`as_handle`	No	If true, cache the fitted result on the server and return result_id + result_uri alongside the JSON payload so a subsequent tools/call can chain without re-running.
`data_path`	Yes	Absolute path or URL to a data file. Supported: .csv / .tsv / .txt (delimited), .parquet / .pq, .feather / .arrow, .xlsx / .xls, .dta (Stata), .json / .jsonl. Schemes: file://, s3://, gs://, https://.
`result_id`	No	Optional handle to a previously-fitted result (returned by an earlier call when as_handle=true). Tools that operate on a fitted object accept this in place of re-supplying data_path + columns.
`covariates`	Yes	Covariate matrix, DataFrame, or column names.
`data_columns`	No	Optional column projection. Parquet/Feather/Stata loaders honour this for fast partial reads.
`data_sample_n`	No	Optional uniform random subsample size (seed=0, deterministic) — useful on huge panels.

Output Schema

TableJSON Schema

Name	Required	Description	Default
No arguments

Schema Changelog

Changes observed during successful MCP inspections.

First observedv0.1.0Sep 15, 2026

TDQS

A4.3/5.0

Behavior4/5

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

Annotations already indicate readOnlyHint=true, so the description's lack of explicit 'read-only' statement is not a gap. The description adds valuable behavioral context: failure modes (extreme propensity scores, super-learner risk), assumptions (unconfoundedness, overlap), and double-robustness property, going beyond what annotations provide. 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.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is moderately long but efficiently structured: it opens with purpose, then validation, assumptions, pre-conditions, failure modes, alternatives, and typical N. Each sentence carries information; no filler. It could be slightly shorter but earns its length.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given that an output schema exists (so return values are documented), the description covers assumptions, failure modes, alternatives, and pre-conditions. It doesn't explicitly describe the estimand parameter's options, but that's in the schema with default ATE. It lacks a note on data format, but that's also in the schema. Overall, it's complete for an agent to call correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so each parameter is documented. The description adds meaning by specifying that treatment must be binary (0/1) and that covariates must comprise the confounding set, which clarifies the semantics of the treat and covariates parameters beyond their schema descriptions. This goes beyond the baseline of 3.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description explicitly states 'Targeted Maximum Likelihood Estimation for ATE/ATT with double-robustness', giving a specific method, estimands, and a key property. It clearly differentiates from siblings by naming alternatives and specifying its focus on ATE/ATT with double robustness, so an agent can distinguish it from sp.dml, sp.aipw, etc.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit alternatives and lists pre-conditions and assumptions (binary treatment, confounding set, n>=500) that guide when to use the tool. It doesn't explicitly state 'use this when X, not when Y', but the alternatives and conditions make the intended use clear.

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

Deploy Server