Application risk intelligence now hinges on identity and supply chain

At a glance

What this is: This is a July 2026 application risk intelligence briefing from Veracode that finds the dominant risk pattern is the convergence of supply-chain compromise, identity abuse, and faster exploitation.

Why it matters: It matters because AppSec and IAM teams now need to govern CI/CD credentials, machine identities, and external exposure as one control plane rather than treating them as separate problems.

By the numbers:

• The top 10 ransomware groups account for approximately 71% of claimed victims.

Context

Application risk intelligence is now less about counting findings and more about understanding which control failures attackers can chain together fastest. In practice, that means identity, supply chain provenance, cloud/IaC posture, and API exposure now sit in the same risk conversation because compromise often begins with access rather than code alone.

The article is especially relevant to IAM, PAM, and NHI programmes because the briefing describes CI/CD credentials, developer tooling, service accounts, and AI-agent credentials as part of the attack surface. That is a material shift from traditional AppSec reporting, where identity was often treated as adjacent rather than central.

Key questions

Q: How should security teams govern credentials used by CI/CD pipelines?

A: Security teams should treat CI/CD credentials as production-grade machine identities and apply least privilege, short lifetimes, and continuous monitoring. The key is to limit what the pipeline can reach, revoke secrets quickly, and audit every privilege path that could let one stolen token spread across repositories, cloud accounts, or SaaS systems.

Q: Why do post-MFA sessions and tokens still create ransomware risk?

A: Because MFA only proves the login event, not the security of what remains after login. If an attacker steals a bearer token, browser session, or delegated grant, they can continue using valid access without repeating the authentication challenge. That is why live session control and rapid revocation matter as much as second-factor policy.

Q: What breaks when application risk scoring ignores identity exposure?

A: Prioritisation becomes detached from exploitability. A flaw with low contextual value may sit ahead of a reachable package path, stolen token, or externally exposed service account that an attacker can use immediately. When identity reachability is missing from scoring, teams overestimate how much their backlog actually reduces risk.

Q: How should organisations respond when AI agents and code assistants touch delivery pipelines?

A: They should govern them as non-human identities with explicit policy boundaries, not as convenience tooling. That means tracking what credentials they can request, what repositories they can reach, and what approvals they can trigger. The practical aim is to stop machine-speed workflows from bypassing the same controls human developers must satisfy.

Technical breakdown

Why CI/CD credentials have become the supply-chain choke point

Modern supply-chain attacks increasingly target the access layer that lets code move through registries, build systems, and deployment pipelines. Once attackers obtain maintainer tokens, OAuth grants, or CI/CD secrets, they no longer need to poison software in a broad, noisy way. They can insert malicious packages, tamper with artefacts, or pivot through trusted automation. That makes provenance and access control inseparable: build integrity fails when the identity attached to the build path is weak, over-scoped, or stolen. Preventing ingress at the credential and approval layer is now more effective than trying to detect every malicious package after publication.

Practical implication: Treat registry and pipeline credentials as high-value identities with explicit lifecycle, scope, and revocation controls.

How identity abuse changes ransomware and extortion paths

The briefing’s identity claim is straightforward: post-MFA session and token theft now underpin much of successful ransomware and extortion activity. That shifts the control question from authentication at login to the security of tokens, browser sessions, device-code flows, and delegated access after authentication has already succeeded. Traditional MFA can still be valid while the attacker’s session remains valid too. In that model, the attacker does not need to defeat the factor again, only to reuse what was already issued. This is why access visibility, token binding, conditional controls, and rapid revocation matter more than point-in-time sign-in checks.

Practical implication: Move monitoring and revocation focus to live sessions, bearer tokens, and delegated access, not only to initial authentication events.

Why context-aware prioritisation is displacing scan-only risk scoring

The report highlights a simple operational problem: vulnerability volume now exceeds human prioritisation capacity, while exploitation windows have compressed to hours. That combination makes flat severity rankings insufficient for application risk intelligence. Organisations need context such as internet exposure, active exploitation, asset criticality, and identity linkage to decide what to fix first. In other words, the control problem is no longer just finding defects, but deciding which ones are immediately exploitable through reachable identities, package paths, or external interfaces. This is where application risk intelligence becomes a governance function, not a scanning function.

Practical implication: Prioritise remediation using exposure, identity reachability, and active exploitation signals rather than CVSS alone.

Threat narrative

Attacker objective: The attacker aims to turn trusted software delivery and identity pathways into a repeatable compromise channel for theft, extortion, and downstream impact.

1. Entry begins through supply-chain access such as stolen CI/CD credentials, maintainer accounts, or delegated tokens that let attackers reach trusted build and packaging workflows.
2. Credential access and abuse then allow malicious packages, poisoned artefacts, or altered dependencies to be inserted through systems that normal trust models treat as approved.
3. Escalation occurs when the compromised path reaches production pipelines, browser sessions, or privileged service accounts, letting the attacker persist inside business-critical workflows.
4. Impact is achieved through ransomware, extortion, code tampering, or downstream compromise of customers and internal environments that consume the tainted software or session state.

Breaches seen in the wild

• LiteLLM PyPI package breach: LiteLLM PyPI supply chain attack, credentials stolen from users.
• reviewdog Action compromise 2025: A stolen maintainer token poisoned reviewdog/action-setup, leaking CI secrets including the tj-actions bot token used in the next attack.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Supply-chain risk has become an identity governance problem, not just a code-integrity problem. When attackers target registries, CI/CD credentials, and maintainer accounts, the control boundary shifts from package scanning to identity lifecycle management. That means the governance issue is ownership, scope, and revocation of the identities that can publish, approve, or deploy software. Practitioners should treat software delivery access as privileged identity infrastructure.

Post-authentication abuse is now the more durable ransomware pattern. The article’s emphasis on session and token theft shows that many programmes still overfit to login events and underfit to what happens after authentication succeeds. In practice, bearer tokens, device-code grants, and delegated sessions create residual access that MFA does not automatically neutralise. The practitioner conclusion is that session governance must sit beside authentication policy.

Application risk intelligence now requires a named concept we can act on: the exposure-to-execution gap. That gap is the time between a vulnerability, credential, or package exposure becoming observable and an attacker turning it into a working intrusion. As the article shows, that window is now measured in hours for some issues, which invalidates slow, scan-led remediation models. Security teams need control decisions that shrink the gap before exploitation, not after detection.

Agentic and AI-assisted attack paths are accelerating the collapse of traditional review assumptions. The report’s references to agentic ransomware and AI-assisted discovery indicate that machine-speed adversaries can now chain identity, tooling, and execution steps faster than manual governance cycles can react. This does not make identity review obsolete, but it does mean review must move upstream into issuance, policy, and runtime control. Practitioners should expect more attacks to arrive as workflow abuse rather than classic malware alone.

Preventive controls are becoming the decisive differentiator in application security maturity. The article links measurable risk reduction to package blocking, contextual prioritisation, and AI-assisted remediation, which is a strong signal that detection-only programmes are losing ground. The governance implication is clear: teams should stop treating prevention as an optimisation and start treating it as the baseline for material risk reduction.

From our research library:

• 92% of organisations expose NHIs to third parties, raising concerns about supply chain security, according to the Ultimate Guide to NHIs.
• Read next: Top 10 NHI Issues

What this signals

Exposure-to-execution gap: application security programmes now need to measure the time between disclosure and usable compromise, not just the number of open findings. When weaponisation happens in hours, governance has to favour exposed assets, active exploit paths, and revoked access over long remediation queues.

The strongest programmes will treat supply-chain ingress, session theft, and agent-issued credentials as one risk surface. That means routing AppSec findings through identity-aware policy, so software delivery controls and access governance reinforce each other rather than operating in parallel.

Veracode’s briefing also reinforces a broader market shift: preventive blocking and contextual prioritisation are becoming board-level expectations because detection alone cannot keep pace with attack velocity. For practitioners, that means the next uplift is less about adding another scanner and more about shrinking the usable attack window.

For practitioners

• Inventory CI/CD and registry identities Map every maintainer account, OAuth grant, pipeline token, and build-service identity that can publish or promote software, then classify them by privilege and business criticality.
• Block malicious packages before ingestion Apply preventive package controls to repositories and developer workflows so newly published, typosquatted, or policy-violating packages are stopped before they reach runners or workstations.
• Tighten post-authentication session controls Focus on token revocation, delegated access review, and session-binding checks for workflows where MFA has already succeeded but access remains live.
• Prioritise exposure-linked remediation Rank findings by internet reachability, active exploitation, and identity exposure so hours-old threats are remediated ahead of stale low-context backlog items.
• Govern AI-generated code and agent credentials Require AI-generated code, AI-assisted fixes, and any agent-issued credentials to pass the same policy gates as human-authored changes before merge or deployment.

Key takeaways

• Application risk intelligence now depends on how quickly attackers can turn trusted software delivery and identity paths into working access.
• Identity abuse, session theft, and supply-chain compromise are converging into the same operational risk pattern for AppSec teams.
• The most effective response is to govern CI/CD identities, block malicious ingress, and prioritise remediation using exposure context.

Key terms

• Device Risk Intelligence: Device risk intelligence is telemetry that assesses whether an endpoint is healthy enough to trust during an identity event. It includes signals such as overlays, malware, remote-access tooling, accessibility abuse, and unusual connectivity, all of which can change the meaning of an otherwise valid transaction.
• Exposure-to-execution gap: The exposure-to-execution gap is the time between when a weakness becomes visible and when an attacker can turn it into a working intrusion. In modern application environments, that gap can be measured in hours, so the control problem becomes reducing usable attack time rather than just finding issues faster.
• Post-Authentication Abuse: Post-authentication abuse happens when an attacker uses valid credentials to perform actions after login rather than breaking authentication itself. For NHI environments, this often means abusing tokens, service accounts, or delegated access to query data, move laterally, or establish persistence while appearing legitimate to basic login controls.
• Supply-Chain Ingress: The entry point through which malicious code, altered dependencies, or compromised identities enter software delivery. It covers package repositories, developer tooling, CI/CD runners, and signing or publishing accounts, all of which can be abused before a defect becomes visible in production.

Deepen your knowledge

NHI Mgmt Group covers identity security, NHI governance, and agentic AI through independent research, practitioner guides, and the NHI Foundation Level course, the industry's only accredited NHI security programme. It is suited to practitioners who need to govern machine identities, service accounts, and AI-agent credentials as part of a broader security programme.