Skip to main content
Glama
vikranthviki

Causal Decision Agent

by vikranthviki
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Related Servers

Alternatives to Causal Decision Agent

No user-submitted related servers found.

    Related Servers

    • jev

      nandansrikrishna
      A
      license
      Not graded
      quality
      C
      maintenance
      Enables agents to evaluate single records or batches with dynamically authored typed questions, returning structured decisions and probabilities.
      14 PyPI
      MIT

    • laya-mcp

      PerryLink
      A
      license
      Not graded
      quality
      A
      maintenance
      Enables typed decisions (yes/no, choice, score) with transparent preflight checks, honest confidence reporting, and health monitoring.
      65 npm
      Apache 2.0

    • Vantage

      abheet19
      A
      license
      Not graded
      quality
      B
      maintenance
      Enables asking product analytics questions in English and receiving answers as real SQL along with the resulting numbers, through read-only MCP tools for funnels, retention, trends, paths, and event counts.
      48 npm
      MIT

    • nightmare-probe-engine

      SweetKenneth
      A
      license
      B
      quality
      C
      maintenance
      Enables generating falsifiable fleet hypotheses, compiling them into minimal bounded probe plans with control cohorts, and evaluating observations to return supported, refuted, or inconclusive verdicts.
      3
      MIT

    • zigmo

      unprimathe
      A
      license
      A
      quality
      C
      maintenance
      Enables AI agents and their users to ask business questions in plain language and receive governed metric values, breakdowns, share-of-total figures, and the exact parameterized SQL behind them, drawn from a single semantic model rather than ad-hoc SQL. It also lets clients inspect available metrics and their definitions, and preview optimized queries without executing them.
      4
      MIT
    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    C2.6/5.0

    Scored across 548 tools

    Disambiguation1/5

    Hundreds of tools are exact or near-exact duplicates: bjs is the same estimator as borusyak_jaravel_spiess and did_imputation; gardner_did duplicates did_2stage; frontdoor duplicates front_door; rosenbaum_bounds and rosenbaum_gamma share identical descriptions; postestimation_contract and postestimation_report are identical; and there are dozens of R-style or article-facing aliases (synthdid_estimate, sc_estimate, did_estimate, xlearner, psm). An agent cannot reliably distinguish the intended tool among so many overlapping aliases and variants.

    Naming Consistency2/5

    Names are mostly snake_case, but the convention is inconsistent: some are concise verbs (did, regress, test, contrast), some are bare nouns (bridge, panel, rate), some are R-package aliases (synthdid_estimate, did_estimate), and some use different spellings of the same concept (frontdoor vs front_door, psm vs psmatch2 vs match). Abbreviated and opaque names like discos, megamma, sqreg, and rlassologit_effects break any predictable verb_noun pattern.

    Tool Count1/5

    548 tools is an extreme mismatch for any server purpose, far beyond even the 50+ upper bound. The surface is a sprawling econometrics library rather than a curated decision-agent toolset, and an agent would face a prohibitive selection problem before doing any actual analysis.

    Completeness5/5

    For the causal-inference domain, coverage is effectively exhaustive: DiD (2x2, staggered, continuous, DDD, event studies), RD (sharp/fuzzy/multi-cutoff/bunching), IV (k-class, weak-instrument, shift-share, MR), synthetic control variants, matching and weighting, mediation, decompositions, sensitivity analysis, survival, time series, CATE/meta-learners, causal discovery, and offline policy learning are all represented. There are no obvious methodological gaps; the problem is surplus, not scarcity.

    Maintenance

    ActivitySlowing
    ResponsivenessNo issues

    Resources