Glasswing and coding agents: are guardrails keeping up?

TL;DR: Frontier models can uncover more than 10,000 critical vulnerabilities in 30 days while enterprise coding agents still hold source code, CI/CD access and secrets, making guardrails too weak to serve as a reliable security boundary, according to Manifold Security research on Anthropic's Glasswing update. The real control problem is assumption collapse: access review and behaviour filters assume a stable, reviewable session, but agentic systems can chain actions, change scope and cause harm inside the same workflow.

Editorial analysis by NHI Mgmt Group, based on content published by Manifold Security: “10,000 Vulnerabilities in 30 Days. Picture This Threat Inside Your Stack.”.

By the numbers:

• The model's true-positive rate on triaged findings sits at 90.6%.
• Anthropic said the program expanded from roughly 50 partners to around 200.
• The update says the program expanded to around 150 additional organizations.

Key questions

Q: What breaks when coding agents have broad access to code, secrets and pipelines?

A: The control that breaks is the assumption that software can be safely trusted with broad standing privilege just because it is running inside a development workflow.

Q: Why do guardrails fail to secure agentic AI workflows?

A: Guardrails fail because they are probabilistic and operate on model output, while the risk lives in the execution chain.

Q: How should security teams set boundaries for coding agents without undermining developer productivity?

A: Treat the agent’s environment as the security boundary.

Practitioner guidance

• Restrict agent authority to task-scoped access Remove broad standing permissions from coding agents and grant only the minimum resources needed for the current task, with explicit separation between read, write and deploy actions.
• Instrument agent activity at the tool layer Log every repository read, shell command, API call and secret access so you can reconstruct the full execution chain rather than relying on prompt logs alone.
• Define behavioural baselines for each agent Establish what normal access patterns look like for each coding agent and alert when it reaches outside its expected files, services or time windows.

Bottom line: Glasswing shows that coding agents can already create meaningful security risk because their authorised access extends into code, secrets and delivery workflows.

Explore further

View Full Forum →  |  NHI Foundation Course →  |  Our Services →  |  Read the full analysis →