TL;DR: SSH bastion workflows can simplify remote administration without storing private keys on servers, but they also preserve a trust model that depends on forwarding, agent handling, and tightly scoped host access, according to StrongDM’s tutorial. The practical lesson is that convenience does not remove identity risk when credentials still move through the session path.
At a glance
What this is: This is a tutorial on SSH agent forwarding through a bastion host, showing how access can move through a trusted session without storing the private key on the server.
Why it matters: It matters because bastion access is still an identity control problem, and IAM and PAM teams need to understand where forwarding, session trust, and host scoping can weaken governance.
Context
SSH bastion access is meant to constrain administrative reach, but the control can still fail if the session path is treated as inherently trusted. The key issue is not whether a private key sits on the server, but whether the access path is governed tightly enough when credentials are forwarded through the session.
In this tutorial, StrongDM focuses on SSH key forwarding through a bastion host and then into an internal guest host. That is a classic privileged access pattern, so the governance question is how to preserve access convenience without expanding the blast radius of a forwarded credential.
Key questions
Q: What breaks when SSH agent forwarding is used without tight bastion scoping?
A: The control breaks when a forwarded session can reach more internal systems than the admin actually needs. At that point, the bastion stops acting as a narrow checkpoint and starts behaving like a privileged relay. The result is broader administrative reach, weaker containment, and a larger blast radius if the session is abused or misrouted.
Q: Why does storing the private key only on the admin machine not eliminate SSH access risk?
A: Because the key location is only one part of the control model. The session still carries delegated trust through the bastion, and that trust can be used to reach internal hosts. Risk persists wherever the reachable set, forwarding rules, or destination permissions are broader than intended.
Q: Where do bastion host workflows most often fail in practice?
A: They fail when the bastion is treated as a convenience layer instead of a governed privilege boundary. Common failure points are overbroad host reach, weak separation between environments, and insufficient review of which internal systems still accept forwarded access from the jump host.
Q: How should teams govern forwarded SSH access in a PAM programme?
A: Teams should govern forwarded SSH access as a session-level privileged path, not as a simple transport detail. That means defining who can use the forwarding flow, which hosts it can reach, and how much internal access the bastion may inherit before the session ends.
Technical breakdown
How SSH agent forwarding changes the trust path
SSH agent forwarding lets a client use a local private key to authenticate through an intermediate bastion without copying the key to the remote system. The remote host receives a forwarded authentication channel, not the key material itself, which reduces direct secret exposure but creates a session-based trust dependency. In practice, the bastion becomes part of the authentication chain, so compromise or misuse of that path can still expose internal reach. The mechanism is useful, but it does not remove the need to govern where the forwarded identity can go once the session exists.
Practical implication: treat forwarded SSH sessions as privileged access paths and scope them as tightly as the destination systems they can reach.
Why bastion hosts still depend on privileged access controls
A bastion host is a controlled jump point that should narrow who can reach internal systems and under what conditions. Even when the private key stays on the admin machine, the bastion still becomes the enforcement point for host reach, network filtering, and session continuity. That means the control surface shifts from credential storage to access path governance. If the bastion is broadly reachable or the internal target accepts too much lateral movement from it, the design weakens into a convenience layer rather than a true privilege boundary. Bastion architecture only works when authorization and network scoping stay aligned.
Practical implication: align bastion network rules, host allowlists, and session controls so the jump host cannot become a generic pathway into the environment.
Where SSH key handling and host scoping can fail
The tutorial’s workflow depends on the private key living on the admin machine and on the bastion only forwarding the authentication context. That arrangement reduces server-side secret storage, but it still leaves several governance questions open: who can use the forwarded agent, how long the session remains trusted, and which internal hosts the bastion can reach. Those are not purely technical details. They are access governance decisions that determine whether the bastion is a controlled administration tool or a shared conduit into sensitive systems. The failure mode is over-trust in the forwarding path.
Practical implication: review bastion host permissions, forwarding rules, and internal destination scope as a single privileged access policy.
NHI Mgmt Group analysis
SSH agent forwarding is an access-path control, not a secret-management cure. The private key may remain off the server, but the session still carries delegated trust from the admin machine to the bastion and then to the target host. That means the real control problem is who can inherit that trust and how far it can travel. Practitioners should treat forwarded SSH access as a privileged delegation chain, not a simple login convenience.
Bastion architecture only works when the jump host has a narrow privilege envelope. If the bastion can reach too many internal systems, the jump point becomes a lateral-movement corridor rather than a guardrail. The article’s workflow shows the benefit of controlled access, but it also shows how quickly a bastion becomes a high-trust intermediary. The governance implication is to define bastion scope as tightly as the workloads behind it.
Forwarded SSH sessions create an identity blast radius that PAM teams have to govern explicitly. The forwarded credential is not static on the server, but the access it enables can still be broad and persistent for the life of the session. That makes the session boundary the meaningful control boundary, especially in environments where admins chain from bastion to guest. PAM programmes should therefore treat forwarding, destination reach, and session duration as one policy domain.
Least privilege in SSH depends on where the trust chain ends, not where the key is stored. The article shows a common misconception: that moving the private key off the server is enough to make access safe. It is not. If internal reach is broader than intended, the effective privilege model still overextends the admin session, so the practitioner must govern the reachable set, not just the credential location.
Named concept: forwarded-session trust debt. This pattern describes the residual risk created when a secure-looking SSH workflow still depends on inherited trust across a bastion chain. The debt shows up when teams assume the absence of key storage equals reduced privilege exposure. In reality, the session itself becomes the asset that needs governance, auditability, and narrow scope.
What this signals
Forwarded-session trust debt: SSH forwarding removes one form of secret exposure, but it creates a different governance problem because the session inherits trust across a jump host. PAM teams should focus on the reachable set and the lifetime of that delegation, not only on where the key is stored.
Bastion hosts are only as strong as the privilege boundary they enforce. If internal reach is broader than the access request that created the session, the architecture shifts from controlled administration to expandable lateral access, which is a governance failure rather than a logging problem.
For practitioners
- Tighten bastion destination scope Limit each bastion host to only the internal systems it actually needs to reach, and separate administrative paths by environment or function so the jump point cannot become a general-purpose relay.
- Treat forwarded sessions as privileged access Require the same review discipline for SSH forwarding paths that you apply to other privileged sessions, including destination limits, approval scope, and logging of the handoff into internal hosts.
- Separate key storage from session authority Keep private keys on the admin device, but do not assume that alone reduces risk. Govern which bastion sessions may inherit the key context and which internal hosts can accept it.
- Audit bastion-to-host reach regularly Review whether the bastion still has access to internal systems that no longer need it, especially where host inventories or team responsibilities have changed since the access path was created.
Key takeaways
- SSH agent forwarding changes where the trust sits, but not whether the session remains privileged.
- A bastion host can narrow access only when its destination scope is tightly governed and regularly reviewed.
- The key control question is not just secret storage, but how much internal reach the forwarded session is allowed to inherit.
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 SP 800-53 Rev 5 and NIST CSF 2.0 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-05 — Overprivileged NHI | Forwarded SSH sessions can overextend access through a bastion host. |
| NHI-10 — Human Use of NHI | The admin’s forwarded SSH workflow relies on a human using a non-human credential path. | |
| Recommendation — Constrain bastion access paths so forwarded SSH sessions cannot inherit broader privilege than needed. Govern human-operated SSH forwarding flows as NHI-backed privileged access, not as ordinary login activity. | ||
| NIST SP 800-53 Rev 5 | AC-6 — Least Privilege | The article’s core issue is controlling how far a bastion session can reach. |
| Recommendation — Apply least privilege to SSH jump paths by restricting which internal hosts each bastion session may access. | ||
| NIST CSF 2.0 | PR.AA-05 — Access Permissions, Entitlements and Authorizations | Bastion forwarding is fundamentally about access permissions and session entitlement. |
| Recommendation — Review permissions and entitlements for bastion forwarding so inherited access stays narrowly authorised. | ||
| MITRE ATT&CK | TA0008 — Lateral Movement | A bastion with broad reach can become a path for lateral movement into internal hosts. |
| Recommendation — Map bastion reach to lateral movement paths and remove unnecessary internal host access. | ||
Key terms
- SSH Agent: An SSH agent is a local process that holds decrypted keys in memory so users do not re-enter passphrases every session. It improves usability, but it also extends the trust boundary to the workstation, which means endpoint hygiene and session control matter.
- Bastion Host: A bastion host is an intermediary system used to reach protected internal resources. It centralises entry but also concentrates trust and availability risk, which means the design only works well when logging, offboarding, and recovery are tightly controlled.
- Privileged Access: Privileged access is any elevated entitlement that can change systems, data, or security settings. When privilege is excessive or poorly scoped, a single compromised identity can create outsized blast radius across environments.
- Forwarded-Session Trust: Forwarded-session trust is the delegated confidence an intermediate host inherits when it passes authentication context to another system. It is a useful pattern for administration, but it expands the control problem to include the whole session path and every host that can receive it.
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.
Published by the NHIMG editorial team on June 8, 2026.
Updated on October 8, 2026.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org