anthropics/defending-code-reference-harness

Skills for threat modeling, scanning, triage, patching, plus an autonomous scanning harness you can /customize

SecurityPython#security

1.1 Steady Magnitude out of 10. Star growth looks organic. Data as of September 14, 2026.

About the project

A reference implementation for autonomous vulnerability discovery and remediation in source code using Claude: skills for threat modeling, scanning, triage, and patching, plus a pipeline with gVisor sandboxing.

Useful for

Build a threat model and run your first static scan with triage
Run the reference pipeline on a C/C++ library to find vulnerabilities
Customize the pipeline for your language or vulnerability class via /customize

README summarized by DeepSeek V4.1 Flash. Details may be inaccurate.

Why it’s trending

10 stars today.
Top new repositories in the last 4 months: #53.

Stars per day

Bars are daily stars, the line is the usual pace. Red marks spike days.

Numbers

Total stars: 7,443
Stars in a day: 10
Forks: 601
Issues and pull requests: 44
Watchers: 44
Language: Python
Created: May 22, 2026
Last push: August 6, 2026

Star trust

Growth looks organic: forks and discussion are in line with active projects, and stars arrive unevenly, the way people give them.

These are heuristics, not a verdict: we judge by the repository’s behavior, not by a list of stargazers.

Spotted in

September 13, 2026Top new repositories in the last 4 months: #52

README badge

Paste it into your README — the badge shows the current magnitude and links to this page.

[![Seismograph](https://gitnova.dev/badge/anthropics/defending-code-reference-harness.svg?lang=en)](https://gitnova.dev/en/r/anthropics/defending-code-reference-harness)

Similar projects

7.0
SnailSploit/Claude-Red
A curated library of drop-in SKILL.md files for the Claude Skills system that primes Claude with offensive security methodology, from SQL injection to EDR evasion and exploit development.

BreakoutSecurityPython+387 stars in a day
6.2
vxcontrol/pentagi
Autonomous multi-agent system in Go for automated penetration testing: AI agents plan and execute steps in an isolated Docker sandbox with 20+ tools (nmap, metasploit, sqlmap).

BreakoutSecurityGo+390 stars in a day
5.0
hezhanleiok/freesub
Automatically tests and publishes subscription pools of free proxy nodes (VLESS, VMess, Trojan, Shadowsocks, Hysteria2, TUIC, AnyTLS), filtering out dead and hijacked nodes via real HTTPS handshakes and TLS validation.

Early signalSecurityPython+50 stars in a day
4.9
tsymbaluyk/maskgate
PII masking service that detects passports, tax IDs, SNILS, bank cards, phones, medical data and secrets in text and files and masks them before they reach ChatGPT, Claude, Gemini. Source code is proprietary; this repo…

Early signalSecurity+44 stars in a day
4.3
zhihui-hu/one-ip
Web tool for IP lookups: identifies datacenter, VPN, Tor and proxy usage, rates reputation from 0 to 100, shows ASN, WHOIS, DNS, CDN, geolocation and checks AI service reachability.

Early signalSecurityTypeScript+17 stars in a day
4.2
realchendahuang/feedsieve
A Chrome extension that flags spam accounts on X (Twitter) with a yellow border and lets you block them manually or in bulk via the native API, synced across all devices.

Early signalSecurityTypeScript+46 stars in a day