Skip to main content
causal-inference-engine Causal inference skill for estimating treatment effects and understanding causal relationships in business data
Aller à l'installation Skills Marketplace Découvrez et explorez les compétences IA créées par la communauté.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Copier le promptAfficher les détails du prompt Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
npx skills add https://github.com/a5c-ai/babysitter --skill causal-inference-engineLa commande reste sur une seule ligne. Faites défiler horizontalement pour la vérifier avant de la copier.
Vous préférez une copie locale ? Téléchargez les fichiers actuellement disponibles dans SkillsMP.
Télécharger Zip Téléchargement... Plus depuis ce dépôt Reference for querying the Atlas knowledge graph through its MCP tools — the SECONDARY enrichment/comparison layer that adds best-practice context to systems you have ALREADY scanned from your real sources (`az`, repos, dirs). Use when you need to look up nodes, edges, kinds, clusters, stats, or wiki pages in Atlas to compare against your real inventory. (atlas graph, query atlas, atlas mcp, search the graph, graph neighbors, atlas record, atlas kinds, enrichment layer)
Atlas turns your STATED NEED into a real systems atlas by SCANNING your actual sources (Azure via `az`, git repos, local dirs) and process/data mining them, THEN enriching against the Atlas knowledge graph. Use this skill when asked to inventory/map your real systems, scan your cloud + repos + directories, mine the real processes or data they contain, or collect their real constraints/gotchas. (atlas, scan my systems, inventory our azure account, map my repos, real systems atlas, process mining, data mining, collect nuances, system discovery)
assimilate-popular-workflows This skill should be used when the user asks to "find skills in the wild", "assimilate popular workflows", "discover SKILL.md files in repos", "research external skills", "find workflow patterns", "survey the skill landscape", "what skills exist out there", or wants to investigate public repositories for extractable processes, babysitter plugins, and reusable procedural insights. Searches GitHub for SKILL.md files, classifies repos by archetype, and maintains structured research under docs/reference-repos/.
Métiers associés SOC
Basé sur la classification professionnelle SOC
name causal-inference-engine description Causal inference skill for estimating treatment effects and understanding causal relationships in business data allowed-tools ["Read","Write","Glob","Grep","Bash"] metadata {"specialization":"decision-intelligence","domain":"business","category":"forecasting","priority":"medium","shared-candidate":true,"tools-libraries":["econml","dowhy","causalml","statsmodels"]} graph {"domains":["domain:business-intelligence"],"skillAreas":["skill-area:statistical-analysis","skill-area:data-analysis","skill-area:quantitative-modeling"],"roles":["role:data-scientist","role:data-analyst","role:research-scientist"]}
Causal Inference Engine
Overview
The Causal Inference Engine skill provides sophisticated methods for estimating causal effects from observational data. It enables business analysts to move beyond correlation to understand true cause-and-effect relationships, supporting evidence-based decision-making for interventions, policy changes, and strategic initiatives.
Capabilities
Propensity score matching
Inverse probability weighting
Difference-in-differences
Instrumental variables
Regression discontinuity
Synthetic control methods
Causal forest implementation
Sensitivity analysis to unobserved confounding
Used By Processes
A/B Testing and Experimentation Framework
Predictive Analytics Implementation
Win/Loss Analysis Program
Usage
Problem Definition
causal_problem = {
"treatment" : "marketing_campaign" ,
"outcome" : "purchase_conversion" ,
"confounders" : ["customer_segment" , "prior_purchases" , "channel" , "region" ],
"instruments" : ["random_assignment_probability" ],
"effect_type" : "ATE" ,
"heterogeneity" : ["customer_segment" , "tenure" ]
}
Propensity Score Matching
psm_config = {
"method" : "propensity_score_matching" ,
"estimator" : "logistic_regression" ,
"matching" : {
"method" : "nearest_neighbor" ,
"caliper" : 0.1 ,
"replacement" : False ,
"ratio" : 1
},
"balance_check" : True ,
"covariates" : ["age" , "income" , "prior_purchases" , "engagement_score" ]
}
Difference-in-Differences
did_config = {
"method" : "difference_in_differences" ,
"treatment_group" : "stores_with_intervention" ,
"control_group" : "stores_without_intervention" ,
"pre_period" : ["2023-01" , "2023-06" ],
"post_period" : ["2023-07" , "2023-12" ],
"parallel_trends_test" : True ,
"fixed_effects" : ["store_id" , "month" ]
}
Causal Forest (Heterogeneous Effects)
causal_forest_config = {
"method" : "causal_forest" ,
"n_trees" : 1000 ,
"honest" : True ,
"effect_modifiers" : ["customer_segment" , "tenure" , "region" ],
"output" : {
"individual_effects" : True ,
"confidence_intervals" : True ,
"variable_importance" : True
}
}
Method Selection Guide Method When to Use Assumptions Propensity Score Selection on observables No unmeasured confounding Difference-in-Differences Pre/post with control group Parallel trends Regression Discontinuity Threshold-based treatment Continuity at threshold Instrumental Variables Unmeasured confounding exists Valid instrument Synthetic Control Aggregate-level intervention Pre-treatment fit Causal Forest Heterogeneous effects Unconfoundedness
Input Schema {
"causal_problem" : {
"treatment" : "string" ,
"outcome" : "string" ,
"confounders" : [ "string" ] ,
"effect_type" : "ATE|ATT|CATE"
} ,
"data" : "dataframe or path" ,
"method_config" : {
"method" : "string" ,
"parameters" : "object"
} ,
"validation" : {
"refutation_tests" : [ "placebo" , "subset" , "random_common_cause" ] ,
"sensitivity_analysis" : "boolean"
}
}
Output Schema {
"effect_estimate" : {
"point_estimate" : "number" ,
"confidence_interval" : [ "number" , "number" ] ,
"p_value" : "number" ,
"standard_error" : "number"
} ,
"heterogeneous_effects" : {
"subgroup" : {
"effect" : "number" ,
"ci" : [ "number" , "number" ]
}
} ,
"diagnostics" : {
"balance_statistics" : "object" ,
"parallel_trends_test" : "object" ,
"first_stage_f_stat" : "number (IV)"
} ,
"refutation_results" : {
"test_name" : {
"original_effect" : "number" ,
"refuted_effect" : "number" ,
"passed" : "boolean"
}
} ,
"sensitivity" : {
"robustness_value" : "number" ,
"interpretation" : "string"
}
}
Best Practices
Clearly articulate the causal question before analysis
Draw a causal diagram (DAG) to identify confounders
Check covariate balance after matching/weighting
Perform sensitivity analysis to unmeasured confounding
Use multiple refutation tests to validate results
Report effect sizes with confidence intervals
Be transparent about assumptions and limitations
Refutation Tests Test What It Checks Placebo Treatment Effect should be zero with random treatment Placebo Outcome Effect should be zero with unrelated outcome Subset Validation Effect should hold in subsamples Random Common Cause Adding random confounder shouldn't change effect
Integration Points
Feeds into Hypothesis Tracker for test results
Connects with Experimentation Manager agent
Supports Predictive Analyst for causal features
Integrates with Bayesian Network Analyzer for causal graphs