← Back to feed

pentagi

GitHub Repo Pretty sure · The hype-to-code ratio is alarming
https://github.com/vxcontrol/pentagi

Penetration testing orchestrator wrapped in so much architectural theater (Neo4j! Grafana! Langfuse!) that it's unclear if the AI actually finds vulnerabilities or just generates plausible-sounding pentesting busywork.

75%
20%
5%
Slop 75%Signal 20%Science 5%

PentAGI claims to be 'AGI for pentesting' but the README is pure feature-list theater: 20+ tools! Neo4j! Knowledge graphs! 10+ LLM providers! Yet nowhere does it show: (1) an actual vulnerability discovered by the AI that a human wouldn't find faster, (2) false positive rates, (3) any benchmark against standard pentesting workflows, or (4) evidence that autonomous orchestration beats running nmap + metasploit yourself. The 'team of specialists' and 'intelligent task planning' are LLM decorato...

11640 stars Go 2026-03-22 439 days old

Become a MFer to rate — log in