Apply Bhaskar's critical realism to analyze phenomena through three ontological domains (real, actual, empirical), identify generative causal mechanisms via retroduction, and examine structure-agency interplay.
SkillScore91Maintained
Automatic
◆Research pack
Last verified 8/27/2026
What it changes
Passed 20 of 22 test cases with the skill, and 14 of 22 without it.
Without skill64%
With skill91%
Checks it passed that the model alone missed
Synthesizes findings into a formal Context-Mechanism-Outcome framework
Includes a section explicitly titled '### Context-Mechanism-Outcome Configuration'
and 6 more in the full report
7 fixed13 passed either way1 still failing1 broke
Pass-rate lift
+27 pp
Tokens
+12%
Agent turns
0%
Response time
-12%
Model gemini-3.6-flashMeasured 8/27/2026v120 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 7
Prompt
I am researching why a community health worker initiative achieved significant reductions in infant mortality in low-income urban informal settlements but completely failed when deployed across rural farming districts. Correlation analysis hasn't explained the divergence. Please conduct a Bhaskarian critical realist evaluation of this case. Structure your answer by detailing what was experienced in the field, mapping out the actual events and environmental conditions, providing a structured matrix of the underlying generative mechanisms (including their causal powers, tendencies, and supporting evidence), synthesizing the context-mechanism-outcome configuration, and drawing out key practical implications for health policy.
What was checked, without and with the skill
Checked
Without skill
With skill
Uses the exact top-level header prefix '## Critical Realist Analysis:' in the output
Missed
Passed
Presents the generative mechanisms table using exact headers Mechanism, Structure/Power, Tendency, and Evidence
Missed
Passed
Includes explicit counterfactual analysis within the actual events evaluation section
Missed
Passed
case-01 · gemini-3.6-flash · 8/27/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md95 linessingle file
Critical Realism (Bhaskar)
Overview
Critical realism posits a stratified ontology: the real (structures and mechanisms that exist whether or not they are activated), the actual (events that occur when mechanisms are activated), and the empirical (events that are observed or experienced). Research must move beyond empirical regularities to identify the generative mechanisms that produce observed phenomena.
When to Use
Moving beyond correlational findings to causal explanation
Designing mixed-methods research that integrates structure and agency
Analyzing why the same mechanism produces different outcomes in different contexts
Critiquing positivist or constructivist approaches as ontologically insufficient
When NOT to Use
When pure prediction without causal understanding is sufficient
When the research question is about subjective meaning-making only (use interpretivism)
When the audience requires strict hypothesis-deductive methodology
Assumptions
IRON LAW: Observable events (empirical) are a SUBSET of what happens
(actual), which is a SUBSET of what exists (real) — research must go
beyond correlations to identify underlying generative mechanisms.
Key assumptions:
Reality exists independently of our knowledge of it (ontological realism)
Our knowledge of reality is always fallible and theory-laden (epistemological relativism)
Some explanations are better than others (judgmental rationality)
Causation is about mechanisms, not constant conjunctions (Humean regularity is insufficient)
Methodology
Step 1: Describe the Phenomenon (Empirical Domain)
Document observable patterns, regularities, and experiences — what do we see happening?
Step 2: Identify Events and Conditions (Actual Domain)
Apply dual-process theory to diagnose whether judgments arise from fast intuitive (System 1) or slow analytical (System 2) processing and identify resulting cognitive biases. Use this skill when the user needs to explain why quick decisions go wrong, design choice architectures that account for cognitive defaults, audit decision processes for heuristic errors, or when they ask 'why do people misjudge probability', 'how to reduce snap-judgment errors', or 'when does intuition fail'.