audit-structural-equivalence

SkillDev tools

Lets your agent test whether a claimed equivalence between two structures really holds and weaken the claim if it doesn't.

Available today. Use it from your connected AI after setup.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the audit-structural-equivalence skill

About this skill

Audit a claimed isomorphism/equivalence by constructing an explicit mapping, testing which structures/operations/invariants are preserved, searching counterexamples, and downgrading the claim to the strongest defensible relation when full isomorphism fails.

What this skill tells your AI

The instructions your AI receives, as published by yogsoth-ai/de-anthropocentric-research-engine in skills/audit-structural-equivalence/SKILL.md and read by ahel’s review.

Purpose

Audit a claimed isomorphism or equivalence and downgrade it to the strongest defensible relation when preservation fails.

Input contract

mode_contracts:
  isomorphism: &structural_audit_input
    required: [source_structure, target_structure, equivalence_claim]
    optional: [required_invariants, operations, constraints]
    constraints: [mapping_domains_and_preservation_obligations_must_be_explicit]
  substructure: *structural_audit_input
  homomorphism: *structural_audit_input
  shared-invariant: *structural_audit_input
  analogy: *structural_audit_input

Execution protocol

Do not perform called SOP operations inline; each loaded SOP owns its contract and thresholds.

  1. You MUST load skill extract-structural-mapping to construct the object, relation, operation, and invariant mapping.
  2. You MUST load skill test-structure-preservation to test preservation. You MUST load skill generate-counterexample to search minimal counterexamples.
  3. You MUST load skill downgrade-equivalence-claim to downgrade the claim and record lost invariants. Deviation: if the claim is only analogy by scope, run the analogy branch and do not report isomorphism.

Output contract

mode_contracts:
  isomorphism: &structural_audit_output
    produces: [claimed_sides, maps_attempted, monsters_found, dimension_count_result, highest_defended_rung, required_wording_change]
    delta_fields: [findings, evidence_updates, uncertainties, decisions]
  substructure: *structural_audit_output
  homomorphism: *structural_audit_output
  shared-invariant: *structural_audit_output
  analogy: *structural_audit_output

Thresholds and quality gates

  • All required relations, operations, and invariants must have pass/fail/unknown status.
  • A full equivalence claim fails on one unprincipled preservation counterexample.

Failure and counterexamples

Do not infer equivalence from surface similarity. Exclusions introduced only after a counterexample are ad hoc unless independently justified.

Provenance map

  • resolved: isomorphism-falsification

Preserved source criteria ledger

sourcesource linekindsource criterion
v4 architecturenode desctextualPreserve structures/operations/invariants; downgrade when full equivalence fails.

Context checkpoint / Delta notes

Append mapping, tested invariants, counterexamples, and final relation.

Mode branches

  • isomorphism: all declared structures and operations must preserve.
  • substructure: test inclusion only.
  • homomorphism: test operation-preserving mapping with relaxed structure.
  • shared-invariant: test named invariant subset.
  • analogy: report transferable relation without equivalence claim.

Signals

GitHub stars
501
Forks
41
Last commit
Sep 2026
Advanced
Catalog kind
skill
Key
audit-structural-equivalence
Source
github.com/yogsoth-ai/de-anthropocentric-research-engine