Sec Triage Agent banner
Bhurva30 Bhurva30

Sec Triage Agent

Security community

Description

Autonomous LLM security triage agent built with Python, Claude Code, Ollama, and Pydantic. Automatically parses SAST/DAST scanner reports to auto-patch hardcoded secrets, trigger Jira/Slack escalations for critical CVEs, and filter out false positives.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

Sec-Triage-Agent: Autonomous Security Triage

Sec-Triage-Agent is an autonomous security orchestration tool designed to automate the initial triage of vulnerability findings from SAST (Static Application Security Testing) and DAST (Dynamic Application Security Testing) tools. By leveraging Large Language Models (LLMs), the agent reduces the manual effort required by security engineers to filter false positives and initiate remediation workflows.

🚀 Features

  • LLM-Powered Analysis: Uses a local or remote LLM (via Ollama) to analyze vulnerability reports and decide on the appropriate triage action.
  • Structured Triage Output: Employs Pydantic to ensure decisions are consistent and machine-readable.
  • Automated Remediation: Integrates with Git to automatically create remediation branches and propose code fixes for low-risk findings.
  • Critical Escalation: Automates the creation of high-priority tickets in SIEM/Jira for critical vulnerabilities.
  • Real-time Alerting: Sends immediate notifications to security channels via Slack for escalated findings.
  • Deterministic Fallback: Includes a rule-based engine to ensure service continuity if the LLM is unavailable.

🏗️ Architecture

The agent operates as a pipeline: `Security Report (JSON)` $\rightarrow$ `Ollama LLM (Analysis)` $\rightarrow$ `Decision Engine` $\rightarrow$ `Tool Execution`

Decision Matrix

Action Condition Resulting Tool Execution
AUTO_FIX Hardcoded secrets / common patterns Git Branch + Patch Suggestion
ESCALATE Infrastructure CVEs / Critical risk Jira Ticket + Slack Notification
REJECT False positives (e.g., test files) Marked as False Positive

📋 Prerequisites

  • Python 3.10+
  • Ollama: Installed and running locally (or accessible via OLLAMA_BASE_URL).
  • LLM Model: qwen2.5-coder:1.5b (default) or other compatible models like llama3.1.

⚙️ Setup Instructions

  1. **Clone the repos