SSO bypass and Golden SAML expose the trust chain problem

At a glance

What this is: NHI Mgmt Group's analysis, drawing on Obsidian Security's blog post: SSO bypass and Golden SAML show how attackers can forge or abuse federated trust to gain access without triggering normal authentication challenges.

Why it matters: It matters because IAM teams often centralise access around SSO, but a weak trust chain can turn one authentication failure into broad application compromise across human and non-human access paths.

Context

SSO bypass is an identity governance problem, not just an application vulnerability. When service providers trust federated assertions without strong validation, attackers can exploit the trust chain instead of stealing passwords or defeating MFA.

The article focuses on how forged tokens, signature validation flaws, and fallback authentication paths undermine the assumptions behind centralized access. For IAM and security teams, the core issue is that valid-looking sessions can exist even when the original authentication never occurred.

Key questions

Q: What breaks when SSO bypass is possible?

A: When SSO bypass is possible, the organisation loses the assurance that authentication events represent real user intent. Attackers can enter with forged assertions, valid-looking tokens, or legacy paths that sit outside normal policy enforcement. That breaks the assumption that MFA and centralised login logs provide complete visibility.

Q: Why do federated trust failures create such broad risk?

A: Federated trust failures matter because one accepted assertion can open many connected applications at once. The risk is not just entry, but inherited access across the downstream ecosystem, which is why service providers must validate assertions independently and not rely on the identity provider alone.

Q: How can teams tell whether SSO bypass controls are actually working?

A: Teams should test whether invalid, unsigned, or misrouted assertions are rejected consistently and whether suspicious post-authentication activity is visible in monitoring. If detection only watches for failed logins, the control is incomplete because bypass attacks can create no failed authentication events at all.

Q: How should security teams handle weak authentication fallback paths?

A: Treat fallback paths as the real control boundary, because attackers often target the weakest recovery option rather than the strongest login factor. Review SMS resets, security questions, and support-assisted recovery first, then remove or harden any path that lowers assurance below your stated authentication standard.

Technical breakdown

Golden SAML and token forgery

Golden SAML works when an attacker obtains the SAML signing key or equivalent trust material and forges assertions that service providers accept as legitimate. Because the forged token is signed correctly, the service provider treats it as proof of authentication even though the identity provider never saw the login. That makes the attack especially hard to catch with traditional authentication logs. The mechanism is not password theft, it is trust theft. Practical implication: Validate the full assertion path, not just the presence of a signed token.

Practical implication: Treat signing material as the trust root and monitor for anomalous post-authentication use, not login failures.

SAML signature validation failures

Signature validation failures happen when service providers accept unsigned assertions, validate the wrong XML element, or allow weak parsing behaviour that separates what is verified from what is consumed. In practice, this means an attacker can present one structure to the verifier and another to the application. The authentication flow appears intact while the authorisation decision is built on corrupted data. This is why SAML implementation detail matters as much as protocol choice. Practical implication: Enforce strict assertion validation and reject any ambiguity in parser handling.

Practical implication: Require strict signature, audience, recipient, and timestamp checks on every federated assertion.

Fallback authentication and shadow access paths

SSO bypass is often amplified by local login routes, legacy protocols, emergency accounts, and abandoned trust paths that sit outside the main identity provider. These paths are dangerous because they preserve access even when central policies are tightened. The article also shows why detection must move beyond authentication events into session behaviour, application access patterns, and lateral movement after access is granted. Practical implication: Inventory every non-SSO path that can still reach production systems.

Practical implication: Remove or tightly govern fallback authentication paths and alert on sessions that do not match normal user behaviour.

Threat narrative

Attacker objective: The attacker wants a cryptographically valid session that bypasses authentication controls and opens access to everything trusted by the federated identity layer.

1. Entry occurs through forged SAML assertions, stolen trust material, or a federated path that lets the attacker appear authenticated without using the normal login flow.
2. Credential access is achieved indirectly because the attacker abuses signing trust or token validation rather than stealing a user password or triggering MFA.
3. Escalation follows when the attacker uses the trusted session to reach multiple connected applications and moves laterally through the SSO trust chain.
4. Impact is broad access to downstream SaaS and cloud applications, often with no failed-login evidence at the point of compromise.

Breaches seen in the wild

• Salesloft OAuth token breach: hackers stole OAuth tokens to access Salesforce data via Salesloft.
• Microsoft Midnight Blizzard breach: Midnight Blizzard (APT29) exploited legacy test account without MFA to breach Microsoft.

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

SSO bypass exposes a trust-chain governance gap, not just a protocol flaw. The problem is that service providers treat federated assertions as proof of identity even when the original authentication never occurred. That means the security boundary sits in the validation logic and trust relationship, not in password policy or MFA coverage. For practitioners, this shifts attention from login friction to trust integrity.

Golden SAML is the clearest example of trust material becoming the real crown jewel. When signing keys or federation trust anchors are compromised, the attacker no longer needs to defeat user authentication. They can mint authority directly, which is why federation keys and certificates belong in the highest protection tier. The implication is that identity programmes must treat federation trust as a privileged asset class.

SSO centralisation creates an identity blast radius that many programmes still under-measure. One successful bypass can open access to dozens of applications, and the blast radius grows when legacy protocols or fallback accounts remain available. Central control only helps if the trust chain is verifiable end to end. Practitioners should measure how many applications still depend on assumptions they cannot independently inspect.

Federated assertion trust: This article names the gap that matters most. The assumption that a signed assertion equals a trustworthy authentication event breaks when attackers can forge, replay, or mislead validation logic. That is a governance failure because the programme is certifying trust in an artifact rather than in the behaviour that produced it. Teams need to rethink how they prove identity continuity across the chain.

Behavioural detection is now part of federation governance. The article makes clear that bypassed authentication often leaves no failed-login trail, so review cycles that depend on authentication logs alone will miss the compromise. Post-authentication anomalies, not just login events, become the decisive signal. For practitioners, the lesson is to align monitoring with where the attacker actually appears: inside the session, not at the door.

From our research library:

• Across one million observed logins, 1 in 4 were password-based rather than SSO, 2 in 5 were not protected by MFA and 1 in 5 used a weak, breached or reused password.

What this signals

Federated trust is the control plane, not a background detail. Once authentication is mediated by trust relationships, the security programme has to verify not just who logged in, but whether the assertion path itself is trustworthy. That is why SSO governance now extends into certificate custody, parser behaviour, and post-authentication telemetry.

Identity teams should think in terms of identity blast radius. The wider the set of applications that depend on one trust chain, the more a single validation weakness becomes a platform-level event. That changes prioritisation from local login hardening to federated boundary assurance and session-level monitoring.

For practitioners

• Audit every federated trust path Map SAML, OAuth, and OIDC trust relationships end to end, including signature validation points, fallback logins, and emergency access routes that sit outside the identity provider.
• Harden assertion validation Require audience, recipient, timestamp, and signature checks on every SAML assertion, and reject any flow that allows optional validation or weak parsing behaviour.
• Inventory legacy and local login paths Identify applications, protocols, and break-glass accounts that still permit access without the primary SSO control, then govern them as explicit exceptions.
• Shift detection to session behaviour Look for unusual application access, impossible travel, rapid lateral movement, and bulk data access after a supposedly normal federated sign-in.
• Protect federation keys as privileged assets Store SAML signing material and related federation secrets under the same governance expectations used for other high-value privileged credentials.

Key takeaways

• SSO bypass works because attackers exploit the trust chain itself, not just user credentials or MFA.
• The highest-value failure mode is forged or misvalidated federated assertions that look legitimate to service providers.
• Security teams need to govern federation keys, fallback access paths, and post-authentication behaviour as a single control surface.

Key terms

• Federated trust chain: A federated trust chain is the set of assumptions linking the identity provider, the assertion or token, and the service provider. In practice, it means downstream applications trust upstream authentication events, so a single weak link can undermine the entire access model.
• Golden SAML: Golden SAML is an attack in which an adversary obtains the SAML signing key or equivalent trust material and forges assertions that service providers accept as legitimate. It bypasses the identity provider's normal authentication flow and can impersonate any user across federated applications.
• SSO Bypass: SSO bypass happens when a user authenticates directly to an application instead of using the organisation’s central identity provider. This breaks visibility, weakens policy enforcement, and can leave security teams blind to who is accessing what. In SaaS environments, repeated bypasses often signal fragmented control and hidden identity risk.
• Audit Validation: Audit validation is the process of proving that governance controls exist and operate consistently enough to satisfy external review. It focuses on evidence, traceability, and repeatability. In identity programmes, validation does not automatically mean risk has fallen, only that the organisation can demonstrate oversight.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are building or maturing an IAM programme, it is worth exploring.