Extract a repeatable empirical pattern from observations, state its support and exceptions, and generalize cautiously without importing an unsupported mechanism.
Installs into .claude/skills of the current project.
Are you the author of Extract Empirical Regularity?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/yogsoth-ai-extract-empirical-regularity)
---
name: extract-empirical-regularity
description: "Extract a repeatable empirical pattern from observations, state its support and exceptions, and generalize cautiously without importing an unsupported mechanism."
---
# extract-empirical-regularity
## Purpose
Extract a repeatable empirical pattern from records while preserving conditions, exceptions, and uncertainty.
## Input contract
```yaml
required: [observations, variable_schema, condition_schema]
optional: [replication_records, measurement_uncertainty]
constraints: [regularity claims require multiple comparable observations or an explicit single-case limitation]
```
## Procedure
1. Normalize observations, units, and conditions.
2. Identify repeated associations, trends, or invariants.
3. Test exceptions, alternative explanations, and measurement artifacts.
4. State the regularity with scope and uncertainty boundaries.
If the regularity is stable enough to express through measurable factors and outcomes, consider `identify-variables` as the next tactic.
## Output contract
```yaml
produces: [regularity_statement, supporting_records, exception_set, scope_conditions]
delta_fields: [findings, evidence_updates, uncertainties, open_questions]
```
## Quality gates
- Conditions and exceptions accompany every regularity.
- Correlation is not labeled mechanism without supporting evidence.
## Failure and counterexamples
Do not generalize a pattern across changed populations or protocols without a comparability check.
## Provenance map
- `resolved: extract-empirical-regularity`
