Webinar Description
Key Takeaways
- Explores the emerging discipline of Agentic Development Security (ADS) for organisations scaling AI-driven software development
- Addresses why traditional application security models struggle with autonomous AI agents that generate and execute code independently
- Covers governance frameworks for maintaining visibility and control when software is created at machine speed
- Designed for CISOs, security architects, DevSecOps professionals and platform engineering leaders
- Examines risks introduced by agent toolchains, autonomous behaviours and AI-generated code in production environments
Introduction
As AI agents transition from coding assistants to autonomous systems capable of invoking tools, executing workflows and generating production-ready code, security teams face a fundamental challenge: traditional application security frameworks were never designed for environments where software creates itself. This webinar from Snyk examines how organisations can maintain security governance without sacrificing the speed and autonomy that make AI-driven development valuable in the first place.
The session introduces Agentic Development Security as a framework for addressing the unique risks that emerge when autonomous systems operate within software development pipelines. For security leaders navigating the rapid adoption of AI coding tools across their organisations, the discussion offers practical approaches to visibility, governance and control in an increasingly automated development landscape.
About This Event
Securing AI-Driven Software Development in the Age of AI Agents is a live virtual webinar hosted by Snyk, featuring expert speakers from the company’s security research and product teams. The session targets application security leaders, platform engineers and DevSecOps professionals working in organisations that have adopted or are scaling AI-driven development practices.
The webinar format allows for focused exploration of emerging security concepts alongside practical implementation guidance, making it suitable for both strategic decision-makers evaluating their security posture and technical practitioners responsible for securing development environments.
Why Traditional Application Security Falls Short
Conventional application security models operate on assumptions that no longer hold in agent-driven environments. Traditional AppSec relies on human developers making deliberate decisions about code, dependencies and architectural choices—decisions that can be reviewed, approved and traced through established workflows. When AI agents autonomously select tools, generate code and execute workflows without human intervention at each step, these foundational assumptions break down.
The challenge extends beyond simply scanning AI-generated code for vulnerabilities. Autonomous agents introduce risks through their tool selections, behavioural patterns and the supply chains that power them. An agent might pull in dependencies, invoke external services or make architectural decisions that would typically require security review—all at machine speed and potentially without visibility into the decision-making process.
This gap between traditional security controls and the reality of agentic development creates blind spots that grow more significant as organisations increase their reliance on AI-assisted coding. Security teams accustomed to reviewing pull requests and scanning repositories find themselves unable to maintain meaningful oversight when code generation happens continuously and autonomously.
Agentic Development Security as a Framework
The webinar introduces Agentic Development Security (ADS) as a conceptual and practical framework for addressing these challenges. Rather than attempting to retrofit traditional security models onto AI-driven workflows, ADS acknowledges the fundamental differences in how autonomous systems operate and proposes security approaches designed specifically for this context.
Central to the ADS framework is the concept of Coding Agent Security, which focuses on real-time governance of agent behaviours and securing the agent supply chain. This involves understanding not just what code an agent produces, but how it produces that code—which tools it invokes, what external resources it accesses and what decision patterns it follows during the development process.
The framework also addresses the need for trust verification at the point of code creation. In traditional development, security scanning typically occurs after code is written and committed. In agentic environments operating at machine speed, this approach creates unacceptable latency between code generation and security validation. ADS proposes mechanisms for ensuring that AI-generated code meets security requirements at the moment of creation rather than through retrospective analysis.
Governance Without Friction
One of the central tensions in securing AI-driven development is maintaining meaningful governance without eliminating the productivity benefits that drive adoption. Organisations implement AI coding tools precisely because they accelerate development velocity—security approaches that reintroduce friction or require human approval for every agent action undermine the fundamental value proposition.
The webinar explores approaches to this balance, examining how organisations can establish guardrails and policies that operate automatically within agent workflows. This includes defining acceptable boundaries for agent behaviour, establishing trust levels for different types of operations and creating escalation paths for actions that exceed predefined risk thresholds.
Visibility emerges as a critical enabler of this governance model. Security teams cannot make informed decisions about agent behaviour without comprehensive insight into what agents are doing, which tools they are using and what code they are generating. The session addresses how organisations can instrument their AI development environments to capture this telemetry without impeding agent performance.
The Agent Supply Chain Challenge
Software supply chain security has become a major concern for organisations over the past several years, with high-profile incidents demonstrating the risks of compromised dependencies and build systems. AI agents introduce a new dimension to this challenge by creating what might be termed an agent supply chain—the tools, models, plugins and integrations that power autonomous coding systems.
When an AI agent selects a tool or invokes an external service, it makes a supply chain decision that carries security implications. Compromised or malicious agent tools could influence code generation in subtle ways that evade traditional detection. The webinar examines how organisations can extend their supply chain security practices to encompass the components that power their AI development infrastructure.
Who Should Attend
This webinar is designed for security and platform leaders responsible for governing AI adoption within software development organisations. The content is particularly relevant for those whose organisations have moved beyond experimental AI coding tool usage into production-scale deployment.
- CISOs and security executives evaluating the risk implications of AI-driven development and seeking frameworks for governance
- Security architects designing controls and policies for environments where AI agents operate autonomously
- DevSecOps professionals integrating security into CI/CD pipelines that include AI code generation
- Platform engineering leaders building internal developer platforms that incorporate AI coding capabilities
- Application security managers adapting their programmes to address AI-generated code at scale
The Broader Industry Context
The emergence of Agentic Development Security reflects a broader maturation in how organisations approach AI adoption. Early enthusiasm for AI coding assistants focused primarily on productivity gains, with security considerations often treated as secondary concerns to be addressed later. As these tools have become embedded in production workflows and evolved toward greater autonomy, the security implications have become impossible to defer.
Regulatory attention to AI systems is also intensifying, with frameworks emerging that may eventually impose specific requirements on how organisations govern AI-generated code. Security leaders who establish robust governance practices now position their organisations to adapt more readily as regulatory expectations crystallise.
For organisations navigating this transition, the webinar offers an opportunity to understand emerging best practices and evaluate how their current security posture aligns with the demands of AI-driven development at scale.