TL;DR: Cloud posture tooling is moving deeper into identity and agentic operations as Prowler’s July 2026 update adds a remote MCP Server for Claude Code, expanded Lighthouse AI workflows, stronger secret detection, one-click AWS onboarding, and new Okta coverage, according to Prowler. The practical issue is not feature breadth but governance: machine-to-machine access, scoped permissions, and auditability now sit at the centre of cloud and identity control.
At a glance
What this is: Prowler’s July 2026 release expands into agentic workflows, secret detection, and Okta scanning, showing how cloud posture tooling is moving deeper into identity governance.
Why it matters: It matters because IAM and cloud teams now have to govern machine-to-machine access, auditability, and identity-layer controls across both cloud and SaaS estate.
By the numbers:
- Prowler now scans Okta with 29 checks mapped to the DISA Okta IDaaS STIG.
Context
Cloud posture tooling is no longer limited to misconfigurations and compliance reporting. In this update, the main governance shift is that remediation, investigation, and identity-layer control are being pulled into the same operating model, which raises the bar for access scope, traceability, and segregation of duties.
That matters for IAM and security teams because agentic workflows can now inspect, prioritise, and help drive remediation across cloud environments while the identity layer becomes part of the control surface. The practical question is no longer whether automation exists, but whether machine-to-machine access is tightly scoped, auditable, and reversible.
Key questions
Q: How should security teams govern AI tools that can act with privileged cloud roles?
A: Security teams should govern AI tools as privileged identity paths, not as harmless application components. That means binding each tool to a task-scoped execution role, tightly controlling who can invoke it, and logging invocation activity as a high-risk event. If a tool can execute actions in a role’s context, it belongs in IAM and PAM review cycles.
Q: Why do machine-to-machine security workflows increase governance risk?
A: Because the workflow can move faster than human review and may act across multiple systems in a single session. If scope, approval, and revocation are weak, delegated access can produce unauthorized changes, widen blast radius, or preserve access after the original need has passed. Governance has to happen at issuance and policy definition, not only after the fact.
Q: What should teams do when their identity provider becomes part of cloud posture management?
A: Bring identity provider settings into the same control review cycle as cloud configuration. That means checking MFA, session policies, token restrictions, and logging as recurring security controls, not one-time admin settings. If the IdP governs access into cloud and SaaS systems, its configuration directly influences containment, resilience, and auditability.
Q: How can organisations tell whether AI-assisted remediation is actually reducing risk?
A: Measure the time from validated finding to safe merge, the percentage of fixes that pass deterministic checks on the first attempt, and the share of high-risk items resolved in the correct owning team. If the AI output is not shortening those cycles, it is only reshaping the queue.
Technical breakdown
Remote MCP access changes how AI tools reach cloud controls
Model Context Protocol, or MCP, creates a standardised way for AI tools to call external systems and act on data. In this release, the important detail is not the interface itself but the trust boundary it creates: Claude Code can reach Prowler Cloud through a remote MCP Server, and that means the governance model must cover tool scope, authentication, and action boundaries. The risk is less about language generation and more about delegated access to operational security workflows. If the access path is not constrained, the agent becomes a privileged operational actor rather than a passive assistant.
Practical implication: constrain MCP tool scopes, log every action, and treat AI-connected security workflows as governed production access.
Agentic triage shifts control from review to issuance time
Prowler’s Lighthouse AI and Claude Code workflow show a broader pattern in cloud security operations: triage is moving closer to the moment of discovery, and sometimes into the same automated workflow that finds the issue. That reduces manual friction, but it also changes where governance has to happen. Instead of relying only on downstream review, teams need controls around what the agent can see, what it can change, and when its actions require human approval. This is especially relevant when findings feed into pull requests or infrastructure-as-code remediation, because the agent’s reach can extend beyond diagnosis into change execution.
Practical implication: define approval thresholds for automated remediation before you let AI-assisted triage touch live environments.
Okta scanning extends cloud posture into identity governance
The new Okta coverage is a reminder that identity provider configuration is part of the cloud security perimeter, not a separate administrative concern. Scanning for MFA enforcement, session policy, password and lockout rules, API token network restrictions, and log streaming connects identity governance to posture management in a concrete way. The critical point is that identity controls now need the same recurring visibility as cloud resource controls. In practice, that means treating the identity layer as a continuously assessed control plane rather than a static admin domain.
Practical implication: bring identity provider settings into the same control review cycle as cloud posture and configuration drift.
Threat narrative
Attacker objective: The objective would be to abuse delegated AI-connected access to alter security findings, trigger unauthorized remediation, or widen control over cloud and identity systems.
- Entry occurs through machine-to-machine access to the cloud posture environment via remote MCP connectivity rather than a human login path.
- Credential or token misuse would let an AI-connected workflow inspect findings, prioritize remediation, or trigger actions outside intended scope.
- Impact would come from over-broad delegated access, where automated remediation changes cloud or identity settings without sufficiently tight governance or human oversight.
Breaches seen in the wild
- Okta support system breach 2023: A support service account credential saved in a personal Google profile let attackers take HAR files and hijack five Okta customers' sessions.
- Cloudflare Thanksgiving breach 2023: One service token and three service accounts left unrotated after the Okta breach gave a nation-state attacker access to Cloudflare's Atlassian systems.
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
Cloud posture tools are becoming identity governors by accident as much as by design. Once an engine can inspect findings, open pull requests, and trigger remediation, it has crossed from observation into operational authority. That is not just a product change, it is a governance change that makes access scoping, audit trails, and exception handling central to cloud security programmes. Practitioners should treat these workflows as controlled change systems, not just smarter dashboards.
MCP creates a new trust boundary that identity teams cannot ignore. Remote tool access for AI systems is only as safe as the authentication, authorisation, and logging wrapped around it. The named concept here is agent-connected control plane drift, where security tooling quietly accumulates delegated power across multiple systems. The practitioner conclusion is straightforward: every AI-connected control path needs explicit ownership and revocation mechanics.
Identity provider posture is now part of cloud blast-radius management. Okta settings such as MFA, session rules, and token restrictions directly shape how far an attacker or misconfiguration can travel. That makes IdP hardening a cloud resilience issue, not only an IAM hygiene issue. Teams should therefore align identity configuration reviews with posture management cadences and incident containment planning.
Secret detection is shifting from exposure hunting to trust validation. Placeholder filtering, live validation, and severity grading matter because not every discovered string is an exploitable secret. The security issue is whether teams can distinguish stale noise from credentials that still authenticate into real systems. Practitioners should optimise for confirmed exposure and rapid revocation, not just larger scan counts.
Agentic remediation only scales when human review moves earlier in the workflow. If the system can propose, prioritise, and enact fixes, the control point has to move upstream to policy, permissions, and change boundaries. That makes this release a useful signal for the wider market: governance is increasingly about constraining machine action, not merely reviewing machine output. Teams should redesign approval paths before automation starts making security decisions at speed.
From our research library:
- 24,008 unique secrets were exposed in MCP configuration files in 2025 alone, the protocol's first year of widespread adoption, according to the State of Secrets Sprawl 2026.
- Read next: NHI Lifecycle Management Guide
What this signals
agent-connected control plane drift: once AI workflows can inspect, prioritise, and change security state, the main governance problem becomes delegated authority rather than model accuracy. Teams should map every AI-connected action path to an owner, an approval boundary, and a revocation process before it reaches production.
The identity layer is now part of blast-radius management, which means posture reviews should include IdP settings alongside cloud configuration drift. That shift makes identity governance operational, not administrative, because session policy, MFA, and token controls now directly influence containment and recovery.
For practitioners
- Define AI tool scopes for security workflows Limit remote MCP and Claude Code access to read-only or tightly bounded actions, and separate investigation privileges from remediation privileges.
- Treat identity provider settings as posture controls Include Okta MFA, session policy, token restrictions, and log streaming in the same recurring review cycle as cloud misconfigurations.
- Build approval gates for automated remediation Require explicit human approval before AI-assisted workflows can open pull requests, apply infrastructure-as-code changes, or trigger fixes in production.
- Prioritise confirmed secrets over placeholder noise Use live validation and revocation playbooks so confirmed credentials are escalated immediately while placeholders stay out of incident queues.
Key takeaways
- Prowler’s July 2026 release shows that cloud posture tooling is moving into delegated action, which makes access scope and auditability the key governance issues.
- The update also extends posture management into identity provider controls, so IAM settings now influence cloud blast radius and recovery planning.
- Teams should separate investigation from remediation, validate confirmed secrets quickly, and review IdP controls on the same cadence as cloud configuration drift.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10 and MITRE ATT&CK address the attack and risk surface, while NIST CSF 2.0 and CSA Cloud Controls Matrix set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-04 — Insecure Authentication | Remote MCP and Okta access hinge on how non-human systems authenticate and are scoped. |
| NHI-02 — Secret Leakage | The release emphasizes secret scanning, live validation, and reduction of exposed credentials. | |
| NHI-05 — Overprivileged NHI | AI-connected workflows and cloud scan roles can overreach if permissions are not tightly bounded. | |
| Recommendation — Restrict NHI authentication paths and verify every delegated access route used by AI-driven workflows. Scan for exposed secrets continuously and revoke confirmed credentials before they can be reused. Limit non-human permissions to the smallest action set needed for investigation or remediation. | ||
| NIST CSF 2.0 | PR.AA-05 — Access Permissions, Entitlements and Authorizations | The article centers on scoping access for AI tools, scan roles, and identity provider controls. |
| Recommendation — Review entitlements for AI-connected and identity-layer services against actual task boundaries. | ||
| MITRE ATT&CK | TA0006; TA0008 — Credential Access; Lateral Movement | Secret exposure and service-account misuse are the main abuse patterns discussed or implied. |
| Recommendation — Map secret leakage and delegated access paths to TA0006 and TA0008 in detection and response planning. | ||
| CSA Cloud Controls Matrix | IAM — Identity and Access Management | The release expands posture management into cloud and identity governance across providers. |
| SEF — Security Incident and Event Management | Audit trails, log streaming, and triage workflows are central to the operating model described. | |
| Recommendation — Use IAM domain controls to keep provider access, session rules, and tokens under continuous review. Integrate security events and audit trails so automated findings remain traceable through remediation. | ||
Key terms
- Remote MCP Server: A Remote MCP server exposes tools over a network so an AI client can discover and invoke them through a standard protocol. In security terms, it becomes part of the identity chain, because it brokers requests, handles authorization, and can expand the blast radius when its OAuth implementation is weak.
- Agentic remediation: Agentic remediation is the automated generation and application of fixes for validated security findings, usually at pull request or pipeline stage. It reduces manual triage pressure by closing routine issues faster, while keeping human oversight for changes that alter business logic or introduce material risk.
- Identity Provider Posture: Identity provider posture is the security state of an organisation’s login and access control layer. It covers session rules, multifactor authentication, network restrictions, tokens, and logging. Because the identity provider often fronts access to many applications, misconfiguration here can create broad exposure even when downstream systems are well protected.
- Delegated Access: Delegated access is permission granted to one identity to act on behalf of another user, service, or system. In NHI environments, this usually appears in OAuth-connected apps and automation tooling. It is powerful, but it must be tightly scoped and reviewed because it can persist long after the original business need ends.
Deepen your knowledge
NHI Foundation Level course, the industry's only accredited NHI security programme, covers NHI governance, machine identity security, and secrets management. It helps security and identity practitioners build the control model needed for delegated access, auditability, and lifecycle governance.
Published by the NHIMG editorial team on August 15, 2026.
Updated on October 12, 2026.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org