AI Governance for developers: A practical framework for deploying autonomous coding agents safely…
summary
Survey of ~100 platform teams shows 78% adoption but 66% lack mature metrics and 25% have no formal policies, so multi-step agents reach only ~39% and levels 4–5 rarely scale. Real incidents include Cursor wiping Pocket OS database and Moonwell losing nearly $2M, which expose the lethal trifecta: private data access, untrusted content, and external comms. Recommended fixes are AIDLC-style pipelines, LLM gateways, distinct agent identities, sandboxed workspaces, and measurable approval gates; Rackspace reports ~25x speedups.
AI-written summary · about
chapters (14)
- 0:00AI governance for developers
- 1:14AI agent adoption in enterprises
- 3:41Governance gaps in scaling agents
- 6:10Security risks from autonomous agents
- 8:18Prompt injection and data poisoning
- 10:45Privilege escalation risks
- 12:10SDLC redesign for AI agents
- 13:14AIDLC success stories
- 15:00Agent deployment failure modes
- 17:07Workspace sprawl governance
- 18:11Four-layer governance stack
- 20:20Agents need governed workspaces
- 21:05Steps to implement agent governance
- 22:31Measuring outcomes and developer experience
related talks
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
36mHow to Safely Deploy AI Agents that Write and Manage your IaC
40mIn AI We Trust? Securing the Future, One Agent at a Time
34mAI Agents: Transforming Autonomous Decision-Making
33mUnlocking Agentic AI: [...]
29mBuilding Trust in AI Agents
26mAI Agents as Your Co-Developer
42mAI for security - friend or foe?