AI agent skill · by @yogsoth-ai

Build Noncircularity Matrix

Cross-tabulate target theory claims/assumptions against validator assumptions and mark independent, shared, derived-from-target, or unknown dependencies.

Archived 2026-10-23 · Source updated 2026-09-29 · License: Apache-2.0. Check the original repository for the latest version.

SKILL.md instructions

# build-noncircularity-matrix
## Purpose
Cross-tabulate target claims/assumptions against validator assumptions and expose circularity.
## Input contract
```yaml
required: [target_claims, target_assumptions, validator_assumptions]
optional: [evidence_channels]
constraints: [each cell is independent/shared/derived-from-target/unknown]
```
## Procedure
1. Enumerate both assumption sets.
2. Populate dependency cells with evidence.
3. Identify copied assumptions and independent failure channels.

If the matrix exposes shared premises or dependent validation paths, consider `detect-pass-by-construction` as the next tactic.

## Output contract
```yaml
produces: [noncircularity_matrix, dependency_summary, independent_channels]
delta_fields: [findings, evidence_updates, uncertainties, decisions]
```
## Quality gates
- Unknown dependencies remain unknown; they are not counted independent.
- At least one non-circular channel is required for validation.
## Failure and counterexamples
A validator that embeds the target claim cannot corroborate it by construction.
## Provenance map
- resolved: circular-validation-audit