TL;DR: Hidden flows of code, internal logic, and proprietary data can leave developer workflows through AI coding assistants, Tumeryk found, while perimeter and DLP controls often miss them. Governance now has to follow the prompts, suggestions, and outputs that cross the IDE boundary.
Editorial analysis by NHI Mgmt Group, based on content published by Tumeryk: “Your Developers are Using AI Code Assistants. Do you have visibility?”.
Key questions
Q: Why do AI coding assistants create new security review risks?
A: They can produce code that looks correct but still contains weak authentication, unsafe data flows, or hidden secret handling mistakes.
Q: What are the signs that release controls are failing around AI-assisted development?
A: Look for missing provenance, repeated contract drift, unreliable rollback decisions, and failures that recur because the root cause was never turned into a reusable gate.
Q: What breaks when AI code assistants are promoted without release governance?
A: Quality failures become operational failures because generated code can compile, deploy, and still violate local contracts, performance constraints, or compliance rules.
Practitioner guidance
- Define approved AI coding use cases Classify which repositories, data types, and development tasks may use external AI assistants, and make the policy visible to engineering teams.
- Inspect prompt and output flows Log, review, and where necessary redact prompts, completions, and plugin calls that can carry source code, secrets, or internal logic.
- Block unapproved plugins and personal accounts Prevent developers from bypassing governance by using unmanaged extensions, personal subscriptions, or unsanctioned model access paths.
Bottom line: AI coding assistants can move sensitive material out of the development environment through ordinary prompts and accepted suggestions, which makes workflow governance a security requirement.
Glossary terms: AI coding assistant, Shadow AI, Egress Governance
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
AI-assisted development now creates a governance boundary inside the IDE, not just inside the security stack. The article is right to treat prompts and generated output as security-relevant data flows, because the development environment has become an egress path for secrets, architecture, and proprietary logic. That changes the identity and access problem as well: the organisation must understand which developer sessions, plugins, and model connections are trusted. The practical conclusion is to govern the developer workflow as a controlled channel, not an informal productivity aid.
A few things that frame the scale:
- 43% of security professionals are concerned about AI systems learning and reproducing sensitive information patterns from codebases, according to the State of Secrets in AppSec.
- Only 44% of developers are reported to follow security best practices for secrets management, exposing a significant developer behaviour gap, according to the State of Secrets in AppSec.
A question worth separating out:
Q: How do organisations govern autonomous coding tools without slowing delivery too much?
A: Organisations should scope the tool, not just the user. Give agents limited edit rights, narrow deployment authority, and auditable checkpoints for multi-file or high-impact changes. That approach preserves speed while preventing autonomous actions from crossing into production without evidence that the change is safe.
👉 Read our full editorial: AI coding assistants expose hidden data flows in developer workflows