Notifications
Clear all
Topic starter
18/08/2026 2:24 pm
TL;DR: ARR grew more than 5x in 2024 and 2025, with H1 2026 on pace to surpass both years as AI-expanded development pipelines and attacker capabilities pushed software supply chain risk upstream, according to StepSecurity. The operational question is no longer whether a single control works, but how to govern trust across developer machines, repos, registries, and CI/CD runtime.
NHIMG editorial — based on content published by StepSecurity: Product 2026 Mid-Year Update
By the numbers:
- AI-related credential leaks surged 81.5% year-over-year in 2025, with the surrounding AI infrastructure leaking 5x faster than core LLM providers.
Questions worth separating out
Q: What breaks when software supply chain controls are only partially automated?
A: Partial automation creates inconsistent enforcement, which means some builds are protected while others slip through manual exceptions or forgotten paths.
Q: Why do AI coding agents increase supply-chain risk?
A: AI coding agents can choose tooling, install packages, and edit lock files in ways that may bypass the controls humans expect in a managed environment.
Q: How do security teams know if a package compromise is actually contained?
A: They should confirm three things: the bad version no longer appears in lockfiles, SBOM records no longer resolve to it, and running services no longer load it.
Practitioner guidance
- Map trust boundaries across the delivery pipeline Inventory where code can be introduced, executed, and promoted across developer machines, repositories, package registries, and CI/CD runners.
- Apply runtime containment to build and release workflows Enable egress blocking, workflow cancellation, and package cooldown controls so a compromised dependency cannot freely phone home or harvest credentials during execution.
- Treat workflow tokens and build secrets as governed NHI assets Assign owners, expiration rules, and revocation paths to tokens and service credentials used in automation, then test whether they can be removed quickly when a package or action is flagged.
What's in the full article
StepSecurity's full blog post covers the operational detail this post intentionally leaves for the source:
- A stage-by-stage view of how Dev Machine Guard, Secure Registry, and Harden-Runner each block compromise in different parts of the pipeline
- Examples of the specific workflow and egress policies used to stop malicious packages from executing or exfiltrating secrets
- The incident search and threat intelligence workflow used to determine whether a compromised component touched developer machines, pull requests, or CI/CD runs
- The product-level release cadence behind the H1 2026 controls and how those releases map to real attack patterns
👉 Read StepSecurity's mid-year update on software supply chain security and AI-driven pipeline risk →
Software supply chain security in 2026: what changed for practitioners?
Explore further
18/08/2026 3:15 pm
AI has turned software supply chain security into a non-human identity problem. The article shows that the control plane now includes packages, runners, workflow tokens, and AI coding agents, all of which behave like machine identities with varying privilege and lifecycle discipline. That means the governance challenge is not just malware prevention, but ownership, scope, and revocation for machine-mediated trust. Practitioners should treat every automated path to code execution as an identity surface.
A question worth separating out:
Q: Who is accountable when a supply chain compromise spreads through trusted credentials?
A: Accountability usually spans release engineering, platform security, and identity governance because the incident crosses multiple trust domains. The practical question is which team owns credential scope, publish rights, and offboarding for automation identities. Frameworks such as NIST CSF and NHI governance models help assign control ownership where a single compromise can affect many systems.
👉 Read our full editorial: AI changed software supply chain risk faster than controls can adapt
