yogthos

yogthos / chiasmus

Public

Chiasmus is an MCP server that gives language models access to formal verification

20
1
100% credibility
Found Apr 09, 2026 at 20 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Chiasmus lets AI assistants use precise math checkers to verify code rules, security flows, and structures with guaranteed correctness.

How It Works

1
🔍 Hear about Chiasmus

While asking your AI helper tough questions about your code's safety or unused parts, it mentions a helper tool for rock-solid answers.

2
📦 Add it simply

You grab the tool with one easy step, like picking up a new gadget for your AI.

3
🔗 Link to your AI

You tell your AI chat buddy to use this new helper whenever needed, and it's ready.

4
💬 Ask smart questions

Now you chat naturally: 'Does user input sneak to the database?' or 'Any forgotten code?' and it digs deep.

5
🔬 Get proven results

Your AI returns clear yes/no answers backed by math proofs, spotting issues no guesswork can find.

🛡️ Code you trust

Your app is verified safe, efficient, and bug-free, giving you total peace of mind.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 20 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is chiasmus?

Chiasmus is a TypeScript MCP server that gives language models access to formal verification tools like Z3 for SMT solving and Tau Prolog, plus tree-sitter-powered code graph analysis. Like the chiasmus definition in rhetoric—a stylistic inversion for emphasis—it flips natural language queries into provably correct logic, verifying RBAC conflicts, dependency resolutions, taint flows, or workflow dead-ends with mathematical certainty. Developers get CLI tools like chiasmus_verify for raw specs, chiasmus_graph for dead code or impact analysis, and chiasmus_formalize for template-guided formalization.

Why is it gaining traction?

It crushes grep-based code tracing: one chiasmus_graph call reveals transitive callers, cycles, or reachability across TypeScript/Python/Go files, no manual chain reconstruction. The template library (chiasmus_skills) covers chiasmus examples like policy contradictions or validation gaps, with LLM filling via MCP—no API keys in Claude Code or Crush setups. Benchmarks show it nails edge cases where LLMs alone falter, delivering unsat cores or concrete models instantly.

Who should use this?

Security engineers auditing access policies or taint paths in microservices. Backend devs resolving package deps or validating workflow states pre-deploy. Refactoring teams on monorepos needing blast radius analysis before changes.

Verdict

Worth a spin for MCP-driven verification in LLM agents—solid docs, benchmarks, and TypeScript polish despite 20 stars and 1.0% credibility score. Early maturity means watch for adapters and Prolog scaling, but it already beats ad-hoc scripts.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.