Agentic Security Pipeline — Security skill for Claude Code
An autonomous agentic pipeline that finds, proves, and patches real C memory-safety vulnerabilities end-to-end using a single 7B open-source LLM (Qwen2.5-Coder) on vLLM, with AddressSanitizer as a gro.
How to install Agentic Security Pipeline
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open FareedKhan-dev/agentic-security-pipeline and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Agentic Security Pipeline does
An autonomous agentic pipeline that finds, proves, and patches real C memory-safety vulnerabilities end-to-end using a single 7B open-source LLM (Qwen2.5-Coder) on vLLM, with AddressSanitizer as a ground-truth oracle.
Alternatives in Security
- Node9 Proxy — The Execution Security Layer for the Agentic Era 209 ★
- Code Audit — Adversarial code review tree 123 ★
- Claudit Sec — Security audit tool for Claude Desktop and Claude Code on macOS — single-command visibility into MCP servers 115 ★
README
🛡️ Agentic Security Pipeline
An autonomous loop that **finds**, **proves**, and **patches** real C memory-safety vulnerabilities — driven end-to-end by a single **7B open-source LLM** on one GPU.
[](LICENSE) [](https://www.python.org/) [](https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct) [](https://github.com/vllm-project/vllm) [](https://github.com/google/sanitizers/wiki/AddressSanitizer) [](secure_code_with_llms.ipynb) [](https://medium.com/@fareedkhandev/building-an-agentic-security-pipeline-that-finds-proves-and-patches-vulnerabilities-6156aeb4b46f)
**Result on the canary target → `3 / 3` bugs discovered · `3 / 3` proven by re-detonation · `3 / 3` patched through a full build → PoC-stop → regression → re-attack ladder · `0` error cells.**
✨ What this is
This repository is a **complete, from-scratch re-implementation of [Anthropic's "defender's loop"](https://www.anthropic.com/research)** — the six-step practice of using LLMs to find and fix vulnerabilities in source code — but with one deliberate twist:
Anthropic drives the loop with Claude. **We drive the exact same loop with an open-source model**, `Qwen2.5-Coder-7B-Instruct`, served locally by **vLLM**. No
Related Skills
Vibe Guard
Full safety check for AI-generated code before pushing. Runs production resilience, security, and comprehensio
Armour
Add Code Armour to this project. Finds and fixes security and legal risks, proves each fix, saves re-tests.
Exploit Chain
Chain multiple findings into a single multi-step exploit. Produces a Foundry test that proves the chain works.
Openclaw Godmode Skill
🚀 GodMode Skill for OpenClaw - Autonomous Claude Code multi-agent coding orchestration. 8 specialized agents
Harness Govern
End-to-end harness governance loop and single front door for the harness- family — measure (harness-budget + h
Claudrunner
Autonomous AI coding agent: finds bugs and security holes, fixes your Jira, Trello, GitHub Issues or Linear ti
Related Agents
Quick Coder
Use for small, focused changes — single function fixes, import updates, renaming, adding a constant. Uses qwen
Incident Investigator
Investigates an LLM-serving incident (vLLM or SGLang) in an isolated context. Reads long runtime logs, runs th
Ark Security Patcher
Fix security vulnerabilities in Ark by researching CVEs, analyzing impact, proposing mitigations, implementing
