Skip to main content
skill CX Codex

proof-theory — Claude Code skill: proof strategies and verification

proof-theory — a skill for mathematical logic proofs and decision trees across strategies. Guides you through direct proof, proof by contradiction, case analysis, and induction; verifies soundness, completeness, and cut elimination.

MIT 1 installs ★ 3,951 updated 2026-01-26

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add parcadei/Continuous-Claude-v3 --skill proof-theory -a claude-code
Open on GitHub

What proof-theory does

Available on Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode. The skill breaks proof strategy into five paths: direct (assume premises, derive conclusion), contradiction (assume negation, reach false), cases (split on disjunction), structural induction (order structures, prove base and step), and verification (check each rule application). Induction requires defining a well-founded ordering, proving base cases, then assuming the property holds for smaller structures to prove it for larger ones.

Verification tools include cut elimination via Gentzen's Hauptsatz (subformulas only, no cuts needed), soundness and completeness checks (provable implies valid; valid implies provable). Commands route through Z3 (z3_solve.py prove) for symbolic reasoning and math_scratchpad.py verify to trace proof dependencies. Both require Bash and file read permissions; commands invoke uv run python -m runtime.harness.

What it helps with

  • choose the right proof method for a logical problem
  • verify each step of a proof and check dependencies
  • eliminate cuts and normalize proofs
  • confirm soundness and completeness of a logical system

Example prompts for the agent

  • Show me the decision tree to pick a proof strategy for this induction problem
  • Verify each step in my proof and flag circular references
  • Apply cut elimination to simplify this derivation

How to install proof-theory

1Skills CLI (Vercel)

npx skills add parcadei/Continuous-Claude-v3 --skill proof-theory -a claude-code

2Manually: put the files in place

~/.claude/skills/proof-theory/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/proof-theory/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about proof-theory

How do I install proof-theory in Claude Code?

Run npx skills add parcadei/Continuous-Claude-v3 --skill proof-theory -a claude-code. The skill installs to ~/.claude/skills/proof-theory/SKILL.md.

What does the proof-theory skill do?

It guides you through proof strategies (direct, contradiction, cases, induction) and validates proofs using Z3 for symbolic reasoning and an internal verifier to check structure, soundness, and completeness.

What do I need to run this skill?

Bash and file read access. Commands execute through uv run python -m runtime.harness with Z3 and math_scratchpad.py as the backend.

Source text

SKILL.md · MIT
---
name: proof-theory
description: "Problem-solving strategies for proof theory in mathematical logic"
allowed-tools: [Bash, Read]
---

# Proof Theory

## When to Use

Use this skill when working on proof-theory problems in mathematical logic.

## Decision Tree


1. **Proof Strategy Selection**
   - Direct proof: assume premises, derive conclusion
   - Proof by contradiction: assume negation, derive false
   - Proof by cases: split on disjunction
   - Induction: base case + inductive step

2. **Structural Induction**
   - Define well-founded ordering on structures
   - Base: prove for minimal elements
   - Step: assume for smaller, prove for current
   - `z3_solve.py prove "induction_principle"`

3. **Cut Elimination**
   - Gentzen's Hauptsatz: cuts can be eliminated
   - Subformula property: only subformulas appear
   - Useful for proof normalization

4. **Completeness/Soundness Check**
   - Soundness: if provable then valid
   - Completeness: if valid then provable
   - `z3_solve.py prove "soundness_theorem"`

5. **Proof Verification**
   - Check each step follows from rules
   - Verify dependencies are satisfied
   - `math_scratchpad.py verify "proof_steps"`


## Tool Commands

### Z3_Induction_Base
```bash
uv run python -m runtime.harness scripts/cc_math/z3_solve.py prove "P(0)"
```

### Z3_Induction_Step
```bash
uv run python -m runtime.harness scripts/cc_math/z3_solve.py prove "ForAll([n], Implies(P(n), P(n+1)))"
```

### Z3_Soundness
```bash
uv run python -m runtime.harness scripts/cc_math/z3_solve.py prove "Implies(derivable(phi), valid(phi))"
```

### Math_Verify
```bash
uv run python -m runtime.harness scripts/cc_math/math_scratchpad.py verify "proof_structure"
```

## Cognitive Tools Reference

See `.claude/skills/math-mode/SKILL.md` for full tool documentation.

© parcadei · MIT · parcadei/Continuous-Claude-v3

Similar skills

All skills for AI and ML →

From the same set161

The whole set →
agentica-server This skill covers running Agentica SDK against a local Claude CLI proxy: startup order, environment variables, and the errors you will hit. CX Codex +1 477installs agentica-sdk A reference for the Agentica SDK: the @agentic decorator, spawn(), agent state, token limits and MCP wiring. CX Codex +1 476installs tdd tdd — a skill that keeps the agent on a strict red-green-refactor loop: failing test first, minimal code next, then verification. CX Codex +1 474installs system-overview system-overview displays the Continuous Claude architecture, showing memory layers, 9 registered hook types, and session continuity mechanisms. CX Codex +1 474installs propositional-logic Guides solving propositional logic problems using truth tables, Z3 proofs, and semantic tableaux. CX Codex +1 1installs predicate-logic A skill for tackling predicate logic in mathematical logic. Handles quantifier analysis, normal form conversion, proof construction, and countermodel search. CX Codex +1 1installs prime-numbers Skill for working with prime numbers in graph and number theory. Offers strategies for primality testing (trial division, Miller-Rabin, AKS), factorization, and applications of Fermat's and Wilson's theorems through Bash and Read tools. CX Codex +1 1installs matrices matrices — a skill for matrix operations: identifies types, performs arithmetic, solves linear systems. Includes decompositions like LU, QR, SVD and works across Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode. CX Codex +1 1installs

Project sponsors

Companies that keep this analytics open