Poolside 2026.04.27

April 27, 2026 On-premises

​

Summary

Manage agents with skills, credentials, secrets, and redaction controls, and gain deeper visibility with improved auditing and tracing. Use Poolside Console and Poolside Chat for faster navigation, agent selection, and search. Automate workflows through the CLI and Identity Management API, and deploy Poolside on more supported platforms. Beta features include plan mode, repositories, and support for third-party model providers.

​

New features

​

Poolside Console

• Skills
  Give agents reusable, task-specific instructions through local skills and organization-managed remote skills. For more information, see Skills.
• Credentials and runtime secrets
  Manage shared credentials in Poolside Console and make them available to agents as runtime secrets. For more information, see Credentials and Secrets.
• Audit log
  Review authentication activity, permission changes, API key usage, and more from the Audit Log page, with configurable retention and S3 export. For more information, see Audit log.
• Redaction patterns
  Define pattern-based rules to redact sensitive values that don’t come from a managed credential or secret. For more information, see Redaction patterns.
• Agent debugging and traceability
  Export trajectory data, inspect time distribution, review inference metrics, and track agent configuration history. For more information, see Agent auditing and Trajectories.
• Product branding
  Customize how Poolside appears in your organization from Settings > Product Branding. For more information, see Product branding.
• Homepage and quick navigation
  A new homepage gives you quick access to common tasks, recent agent activity, and console-wide search.
• Model connection testing
  Test a model configuration before saving it using Test Connection.

​

Poolside Assistant (IDE)

• More flexible sandbox configuration
  Configure sandboxes with volume mounts, workspace-specific settings, and a more detailed selector so you can reuse local caches and configuration files while keeping projects isolated.
• Sandbox session restoration
  Restore sandbox sessions when you switch conversations or restart your IDE, so you do not need to rebuild environments for related work.
• Auto-allow mode and sandbox controls
  Administrators can grant the Auto Approve Commands permission to let trusted workflows run with fewer prompts, with optional sandboxing for stronger runtime boundaries. For more information, see Tool permissions and Sandboxes.

​

Poolside Chat

• Poolside Chat replaces Poolside Web Assistant
  Poolside Chat is the new name for the experience previously called Poolside Web Assistant.
• Search in Poolside Chat
  Search for previous conversations from the sidebar or with Command+K (macOS) or Ctrl+K (Windows and Linux).

​

Poolside Agent CLI

• Interactive mode
  Work with agents in the terminal by running pool, with support for slash commands such as /skills and trajectory review after a run. For more information, see Interactive mode.
• Automated mode
  Run one-shot agent workflows with pool exec, including unattended runs with --unsafe-auto-allow. For more information, see Automated mode.
• ACP-compatible editor support
  Use agents from JetBrains, Zed, Neovim, and other ACP-compatible editors through pool acp. For more information, see Editor integration (ACP).

​

API

• Identity Management API
  Manage users and team assignments programmatically for user lifecycle and provisioning workflows. For more information, see Identity Management API.

​

Platform and infrastructure

• Additional deployment options
  This release adds support for Red Hat OpenShift, upstream Kubernetes, and AWS-native deployment options. For more information, see Supported configurations.
• Improved model serving
  Run model deployments with improved reconciliation, authenticated serving connections, and direct stop-token handling.

---

​

Beta features

• Plan mode
  Review an implementation approach before the agent changes your code. For more information, see Plan mode.
• Repositories
  Connect shared files from Git repositories, S3 buckets, or uploaded archives for remote skills and reference content. For more information, see Repositories.
• Third-party model providers
  Integrate any model in your experience. Leverage all the security and compliance features available in Poolside Console with any model. For more information, see Models.

---

​

Requirement changes

• Visual Studio 17.9 is now required
  Poolside Assistant for Visual Studio now requires Visual Studio 17.9 or later.

---

​

Breaking changes

• Existing pool flags move to pool exec
pool now opens interactive mode, and non-interactive commands now use pool exec. The flags are unchanged. Update any scripts, aliases, or command snippets by replacing pool with pool exec.

  Before	After
  pool -p "Summarize this repo"	pool exec -p "Summarize this repo"
  pool -f prompt.txt	pool exec -f prompt.txt
  pool < prompt.txt	pool exec < prompt.txt

---

​

Known issues

​

Poolside Console

• Multiple open Poolside Console tabs may cause authorization issues
  Opening the Poolside Console in multiple browser tabs simultaneously may result in authorization errors.
  Workaround: Use a single tab when working in the Poolside Console.

​

Poolside Assistant (IDE)

• Apply Changes may fail after multiple edits in sandboxed sessions
  In some cases, applying changes from a read-only sandbox may fail after follow-up edits in the same conversation.
  Workaround: Apply changes at the end of a conversation before starting a new one.
• Binary files may appear in diffs in sandboxed sessions
  Binary files created in read-only mode may be incorrectly rendered or included in line-count diffs during change review.
  Workaround: Ask the agent to remove built binaries before using Apply Changes, or add them to .gitignore.
• Limited tooling in default sandbox images
  The bundled default images provide a minimal toolset and may not include commonly expected development tools.
• Private container registries require pre-authentication
  Poolside Assistant cannot negotiate container registry authentication. Runtimes must already be authenticated to pull from private registries.

​

Platform

• Trajectory storage growth
  All execution trajectories are stored in the database, which can lead to rapid database growth over time.
  Workaround: A pruning script is available for administrators to manage storage usage.

​

Component versions