Hedgemony
A free Python tool that catches AI-generated code hallucinations — invented packages, fabricated APIs, wrong call signatures — using deterministic analysis instead of asking another AI model to check.
🔗 Visit Hedgemony
Description
The usual way to catch mistakes in AI-generated code is to ask another AI model to review it — but that just adds a second layer of guessing on top of the first. Hedgemony instead checks deterministically: does this package actually exist, does this method actually exist, does this function call actually match its real signature.
Hedgemony runs a two-pass analysis — static checks without executing code, plus optional sandboxed dynamic testing with CPU/memory/process limits — across six error categories: nonexistent packages, wrong module paths, bad imports, invalid attributes, wrong keyword arguments, and incorrect argument counts. It also verifies that docstring examples actually match real behavior, works entirely from the Python standard library with zero external dependencies, and outputs to terminal, Markdown, HTML, or JSON. A `--board` flag lets you compare multiple models' outputs, and a `--no-run` safety mode disables dynamic execution entirely. It's free and open source under SSPL-1.0.
💬 Our review
The short version: Hedgemony's core idea — verify AI-generated code deterministically rather than asking another LLM to judge it — is a genuinely sound approach, since an LLM reviewer can hallucinate its own approval just as easily as the original code hallucinated an API.
Tools like CodeRabbit or Codacy do broader code review but rely on either static-analysis heuristics or another model's judgment, not the specific package/API-existence verification Hedgemony targets; standard linters like mypy or pylint catch type and style issues but generally don't verify that an imported package or method actually exists in the installed environment. The narrowest, most direct comparison is simply not using anything and trusting the AI-generated code as-is — which is the actual status quo for a lot of agentic coding workflows, and exactly the risk Hedgemony is built to catch before it ships. Being dependency-free, sandboxed for dynamic checks, and free makes it low-friction to add as a guardrail step after any AI code-generation pass.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Open source sous licence SSPL-1.0, aucun coût
Pros
Vérification déterministe, pas besoin d'un second modèle IA
Zéro dépendance externe (stdlib Python uniquement)
Analyse statique et dynamique sandboxée
Six catégories d'erreurs couvertes, formats de sortie multiples
Cons
Limité à Python pour l'analyse dynamique
Licence SSPL-1.0 potentiellement restrictive pour du SaaS
Projet jeune, communauté encore réduite