detect-contradiction
SkillDev toolsLets your agent find and classify contradictions between claims, derivations, or pieces of evidence.
Available today. Use it from your connected AI after setup.
No other account needed.
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 detect-contradiction skill
About this skill
Detect an explicit contradiction in logical derivations, claims, or evidence relations; classify whether it is formal, empirical, scope-dependent, or unresolved.
What this skill tells your AI
The instructions your AI receives, as published by yogsoth-ai/de-anthropocentric-research-engine in skills/detect-contradiction/SKILL.md and read by ahel’s review.
Purpose
Detect and classify explicit contradictions in derivations, claims, or evidence relations.
Input contract
required: [claims_or_derivation, evidence_relations, scope_context]
optional: [logic_rules, causal_graph, contradiction_taxonomy]
constraints: [distinguish formal, empirical, scope-dependent, and unresolved contradiction; preserve both sides]
Procedure
- Normalize claims, predicates, scope, and evidence polarity.
- Compare opposing claims or support/contradict relations under shared scope.
- Classify the conflict and record the exact conflicting statements.
- Create a contradiction record and identify the evidence needed for adjudication.
If the contradiction concerns treatment, component, or effect attribution, consider analyze-intervention as the next tactic.
If the contradiction rests on a shared but untested premise, consider challenge-assumption as the next tactic.
Output contract
produces: [contradiction_register, conflicting_relations, classification, adjudication_questions]
delta_fields: [findings, evidence_updates, uncertainties, open_questions]
Quality gates
- Support and contradict evidence are both recorded; contradiction cannot be hidden by selecting one polarity.
- Scope mismatch is labeled scope-dependent rather than formal contradiction.
- Each contradiction includes both endpoints and a reproducible comparison basis.
Parameterization
The caller must provide claim schema, evidence polarity vocabulary, scope rules, logic/causal relation types, and the adjudication status vocabulary.
Failure and counterexamples
Reject conflicts caused only by incomparable metrics, different scopes, or duplicate text; retain them as unresolved comparability issues.
Provenance map
- resolved: stress-test/contradiction-detection
- resolved: knowledge-structuring/contradiction-flagging
- concept: stress-test/detect-contradiction
- intermediate: Pass3/detect-contradiction
- intermediate: Pass3/flag-contradictory-evidence
Preserved source criteria ledger
| source | physical line | kind | source criterion |
|---|---|---|---|
| knowledge-structuring/contradiction-flagging | 18 | gate | Record simultaneous support and contradict evidence and do not ignore either. |
Signals
- GitHub stars
- 501
- Forks
- 41
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Key
detect-contradiction- Source
- github.com/yogsoth-ai/de-anthropocentric-research-engine