ssh-access-manager

Manage SSH sidecar access routing to workbench target containers via ForceCommand.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/spigell/my-agents --skill ssh-access-manager
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ssh-access-manager
Source: https://github.com/spigell/my-agents/tree/main/.agents/skills/ssh-access-manager
Command: npx skills add https://github.com/spigell/my-agents --skill ssh-access-manager

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It removes friction and mistakes in how workbench users reach the correct target containers over SSH sidecars, especially when ForceCommand routing, zmx menu/profile attachment, and in-cluster kubectl hops must stay deterministic.

Core Features & Use Cases

  • SSH sidecar ForceCommand wiring: Ensures SSH requests are routed through the correct wrapper entrypoint with predictable behavior.
  • Wrapper-based container hop & guardrails: Uses helper logic to move from the SSH sidecar to the appropriate target container and keeps non-interactive passthrough consistent via SSH_ORIGINAL_COMMAND semantics.
  • Deterministic zmx profile and shared-session conventions: Standardizes profile naming, shared attach behavior, and shortcut attachment rules so operators land in the intended session.
  • Use Case: When multiple workbench agent kinds (codex, gemini, gemini-secondary, claude) must be accessed from the same SSH sidecar, this skill guarantees the correct container and session/profile are selected and attached without relying on ad-hoc operator steps.

Quick Start

Use the ssh-access-manager skill to set up SSH sidecar ForceCommand routing for a workbench agent kind by ensuring the sidecar runs prepare-ssh-config with the desired agent-kind.

Frequently Asked Questions about ssh-access-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I route SSH access to a specific workbench container through a Kubernetes sidecar?

Route SSH access by configuring ForceCommand on the sidecar to intercept requests and use a wrapper to hop into the correct target container via kubectl exec. This ensures deterministic routing without relying on manual operator steps.

Why does my non-interactive SSH command fail when connecting through a sidecar to a workbench pod?

Non-interactive SSH commands fail when ForceCommand routing does not properly preserve SSH_ORIGINAL_COMMAND semantics. The wrapper must pass through the original command directly to the target container instead of dropping or altering it during the hop.

Can I access multiple workbench agent kinds like codex, gemini, and claude from the same SSH sidecar?

Yes, you can access multiple agent kinds from the same sidecar by running prepare-ssh-config with the desired agent-kind. This standardizes profile naming and session attachment so each request reaches the correct container deterministically.

How do I standardize zmx profile naming and shared-session attach behavior for workbench containers?

Standardize zmx profiles by enforcing deterministic naming conventions and shared-session attach rules within the SSH sidecar configuration. This guarantees operators land in the intended session without shortcut indirection causing ambiguity.

What's the best way to prevent shortcut indirection when attaching to a shared workbench session over SSH?

Prevent shortcut indirection by enforcing direct carrier attach rules within the ForceCommand wrapper logic. The wrapper validates Bash syntax usage patterns and centralizes attach logic, ensuring the session connection is direct and predictable.

Does ssh-access-manager require any external dependencies or installed components to manage workbench SSH routing?

No external dependencies or components are required. The skill operates by validating Bash syntax usage patterns and centralizing the routing and attach logic natively through the existing SSH sidecar and kubectl exec infrastructure.