3 ms·
Show HN: Phantom Guard – Detect AI-hallucinated package attacks
Phantom Guard is a CLI tool that catches "slopsquatting" attacks before they compromise your supply chain.
The attack vector: AI assistants hallucinate package names → attackers register those names with malware → developers install malware thinking it's legit.
How it works:
1. Checks if packages exist on registries
2. Matches against 10+ AI hallucination patterns
3. Detects typosquats of top 3000 packages
4. Analyzes metadata (age, downloads, maintainers)
```
pip install phantom-guard
phantom-guard validate flask-gpt-helper
# HIGH_RISK: Package not found, matches pattern
```
Performance: <10ms cached, <200ms uncached.
Try the live demo: https://matte1782.github.io/phantom_guard/ https://matte1782.github.io/phantom_guard/
GitHub: https://github.com/matte1782/phantom_guard https://github.com/matte1782/phantom_guard
- dmarwicke 9mo agodoes this end up flagging legit packages that just have 'ai' or 'gpt' in the name? feels like half of pypi would trigger at this point
- matteo1782 9mo agoGreat question! No, Phantom Guard won't flag legit packages like openai, langchain-openai, or gpt-engineer. The primary signal is whether the package exists on the registry. We query PyPI/npm directly: - If a package exists → it gets a low/safe risk score - If a package doesn't exist → that's the main red flag for slopsquatting Pattern matching (like AI-related terms) is just one of many weighted signals, and it's far outweighed by existence. In fact, popular packages get a negative weight that actively reduces their risk score. The attack we're detecting is when an LLM hallucinates a package name like flask-gpt-utils that sounds plausible but doesn't exist. A real attacker could then register that name and wait for developers to pip install it. We test against the top 1000 PyPI packages and target <5% false positive rate. If you're importing openai or transformers, you're fine.