Causal Decision Agent

Overview Schema Source Related Servers Discussions Admin

mc_panel

Read-only

Estimate treatment effects from panel data using matrix completion, appropriate when a substantial control block is present.

Instructions

Estimate treatment effects using matrix completion. Validation: validated evidence tier (known-truth, reference, external-parity, or Monte Carlo artifact). Do NOT use when: the panel is nearly fully treated -- matrix completion needs a substantial observed-control block to recover the low-rank structure. Cost: Iterative soft-impute: one SVD of the N x T outcome matrix per iteration, i.e. O(max_iter x N x T x min(N,T)). n_bootstrap multiplies the whole loop -- this is the dominant cost on wide panels. Assumptions: The relationship between the treated unit and controls is stable absent the intervention (causal_impact); Matrix-completion: the untreated potential outcomes follow a low-rank factor structure with treatment as the missingness pattern; No concurrent intervention affects the controls. Pre-conditions: A treated unit with a pre-period and a set of control series, or a panel with a low-rank structure. Failure modes: Pre-period fit is poor or controls are themselves affected by the intervention -> Re-select controls, lengthen the pre-period, or use synthetic-control / DiD diagnostics. Alternatives: sp.synth, sp.sdid, sp.gsynth. Typical minimum N: 50.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`y`	Yes	Outcome variable.
`tol`	No	Convergence tolerance.
`time`	Yes	Time period variable.
`unit`	Yes	Unit identifier variable.
`alpha`	No	Significance level.
`treat`	Yes	Binary treatment indicator (0/1). Can be staggered.
`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
`max_iter`	No	Maximum iterations for the proximal gradient algorithm.
`max_rank`	No	Maximum rank for the completed matrix. If None, no constraint.
`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.
`lambda_reg`	No	Nuclear norm regularisation parameter. If None, estimated via the universal threshold: lambda = sigma * sqrt(n).
`n_bootstrap`	No	Bootstrap iterations for standard errors.
`data_columns`	No	Optional column projection. Parquet/Feather/Stata loaders honour this for fast partial reads.
`random_state`	No	Random seed or RandomState for reproducible stochastic steps.
`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

A4.5/5.0

Behavior5/5

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

Annotations already declare readOnlyHint=true and openWorldHint=false, so the safety profile is covered. The description adds substantial behavioral context beyond that: computational complexity (one SVD per iteration, n_bootstrap multiplying the loop), key assumptions (low-rank factor structure, stable pre-treatment relationship absent intervention), and failure modes with consequences. 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 long but every labeled block (Validation, Cost, Assumptions, Pre-conditions, Failure modes, Alternatives, Typical minimum N) earns its place and is scannable for an LLM. The core purpose is front-loaded and no sentence is tautological or redundant; only the density of the assumption list keeps it from a 5.

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

Completeness5/5

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

For a complex estimator with 17 parameters and an output schema, the description is close to complete: it covers purpose, when-not-to-use, cost, assumptions, preconditions, failure-mode remediation, alternatives, and minimum-N guidance. Since an output schema exists, the description need not explain return values.

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 the baseline is 3. The description adds performance semantics for parameters: n_bootstrap 'multiplies the whole loop... this is the dominant cost on wide panels' and max_iter appears in the per-iteration SVD cost, which is genuinely useful for resource planning. Required columns (time, treat, unit, y) are self-explanatory and already documented in the schema.

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

Purpose4/5

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

The description opens with a specific verb+resource: 'Estimate treatment effects using matrix completion.' It is scoped by an explicit exclusion condition (nearly fully treated panels) and names alternatives (sp.synth, sp.sdid, sp.gsynth), though it does not explicitly differentiate from the closely named siblings 'matrix_completion' and 'mc_synth' in the tool list.

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

Usage Guidelines5/5

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

Usage guidance is explicit and complete: 'Do NOT use when: the panel is nearly fully treated' states a concrete exclusion, 'Alternatives: sp.synth, sp.sdid, sp.gsynth' routes the agent to substitutes, and 'Pre-conditions' plus 'Failure modes' with remediation ('Re-select controls, lengthen the pre-period') finish the selection story.

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

Deploy Server