July 2026

Define Agents for your team's workflows

30 July release

An agent is an organization-scoped assistant with its own instructions, example prompts, and bindings to Venue tools, MCP server tools, and other agents.

• Organization registry - Browse and search agents from Agents in the sidebar (under AI).
• Custom instructions - Write what the agent does and how it should respond, in prose.
• Capability bindings - Grant access to Venue tools, individual MCP server tools, or other agents. Use Add all to bind every tool from a server at once.
• Templates - Start from a predefined template to pre-fill name, description, instructions, and example prompts.
• Safety controls - Require human approval before the agent executes sensitive actions.
• Lifecycle management - Move agents through Draft, Ready, Disabled, and Archived states. Duplicate, disable, or archive from the detail page.

Admins and Editors can create agents. Contributors, Viewers, and Demo users can browse them. The Activity Log records every create, update, and delete.

See Agents for setup details and example configurations such as an Atlassian-aware agent or a Kubernetes and Helm troubleshooting agent.

Select the right Agent in the Assistant

30 July release

The Assistant now supports agent selection alongside the general Venue.sh assistant.

• Agent selector - Choose a ready agent from the Assistant header before or during a conversation.
• Scoped tools - Select an agent and the Assistant uses that agent's instructions and only the tools bound to it, including full MCP server tool access where configured.
• Page context on AI pages - On supported AI pages (e.g., Agents, MCP Servers, Skills), use Add current page to context so the Assistant knows what you are looking at.

For example, select a DevOps agent and ask:

• "Why is the payments-api pod crashing in staging?"
• "Show me the Helm values for the checkout-service release in production."

Or ask the general assistant about your agent catalog:

• "What agents are available in our organization?"
• "Which agents have access to Jira tools?"

Ask the Assistant about Development Insights

30 July release

The Assistant reads the same Grafana panel data you see in the Development Insights UI, so its answers match your dashboards.

Three read-only tools cover the dashboards most teams use for delivery health:

• DORA - Deployment frequency, lead time for changes, change failure rate, and recovery time.
• Engineering Overview - Cross-provider delivery health across Jira and Git.
• Engineering Throughput & Cycle - Throughput, quality, and cycle-time workflow metrics.

Ask a delivery question and the Assistant calls the matching tool, applies any filters you mentioned such as project, time range, or DORA report year, then summarizes the panel data it gets back. It defaults to the last six months if you do not name a range.

Access follows your existing permissions. You need Development Insights read permission, and Development Insights enabled for your organization.

Try asking:

• "What is our deployment frequency this quarter?"
• "How does our lead time compare to last month?"
• "Summarize throughput and cycle time for the platform team."

Provider-specific Git dashboards (GitLab, GitHub, Bitbucket) are not available through Assistant tools in this release. Use the dashboard UI for those views.

Updated Assistant model lineup

30 July update

The built-in model catalog adds Claude Opus 4.8, Claude Sonnet 5, and GPT-5.5 Pro. Your Venue administrator controls which models appear in the Assistant model dropdown through Settings → Assistant.

The four DORA metrics in one view

28 July release

A table at the top of the dashboard places each of your metrics in its benchmark tier: low, medium, high, or elite. Below it, a stat panel per metric shows your current value for the selected time period.

• Deployment Frequency - How often your organization deploys code to production, measured in deployment days per week or month.
• Lead Time for Changes - The median time for a code change to reach production, from first commit to deployment.
• Change Failure Rate - The percentage of production deployments that result in incidents.
• Failed Deployment Recovery Time - The median time to recover when a production deployment causes an incident.

Tier colors run purple for elite, green for high, yellow for medium, and red for low.

Benchmarks you can choose

28 July release

DORA's research benchmarks changed between report years, and the dashboard supports both:

• 2023 report - The current benchmark thresholds, including the Failed Deployment Recovery Time metric.
• 2021 report - The earlier thresholds, which measure Median Time to Restore Service instead.

Use the DORA Report selector at the top of the dashboard to switch between them. Panel titles and benchmark thresholds update to match the report you pick.

Trends and drill-downs

28 July release

Every metric comes with history and a way into the raw data.

• Monthly trend charts - Bar charts plot deployments per month, median lead time, change failure rate, and recovery time over time.
• Drill-down dashboards - Click any metric to open a dedicated view of the underlying deployments, pull requests, or incidents.
• Project and time filters - Use the Project filter to scope metrics to one or more projects, and the date range selector to adjust the window. The dashboard defaults to the last six months.

Enterprise SSO from Settings

27 July release

Organization admins on Enterprise plans can now configure their own SSO provider from Settings, without waiting on manual setup.

• Self-serve configuration - Connect your identity provider and manage the connection from within Venue.sh.
• Fewer support round-trips - Admins no longer need to coordinate with the Venue.sh team to stand up SSO for their organization.

Navigate Venue.sh without touching the mouse

27 July release

Press ⌘K on Mac (or Ctrl+K elsewhere) to open global search and jump straight to any sidebar destination.

• Search everywhere - Type a page name or common alias to find the destination you want, wherever you are in the dashboard.
• Keyboard-first - Move through results with arrow keys and Enter.
• Respects your access - Only destinations you have permission to see appear in results.

Development Insights filter improvements

27 July release

Filtering Development Insights dashboards is more predictable.

• MultiSelect filters with "All" - Project, priority, and repository filters use checkbox multi-selects with a dedicated All option, replacing the previous single/multi toggle.
• Clearer date range control - The Clear button for the date range filter now sits next to its label, so it stays visible when the label wraps on smaller screens.

Assistant sidebar polish

Ask from any page

13 July release

Open the right-side sidebar while you browse your catalog, review TechDocs, launch a template, or work in another supported area of the dashboard.

• Start anywhere - Begin a new chat or return to your current conversation without navigating to the full Assistant page.
• Add page context - On a supported page, choose Add current page to context to tell the Assistant what you're looking at.
• Continue past conversations - Search your chat history, resume a conversation, pin important chats, or delete conversations you no longer need.
• Stay in control - Choose an AI model and thinking effort, enable web search where available, or stop a response mid-stream.
• Switch to the full experience - Open a sidebar conversation on the full Assistant page when you want more room.

For example, add the current page to context and ask:

• "Summarize this catalog entity"
• "What inputs do I need before running this template?"
• "What repositories need attention here?"

More reliable repository ingestion

13 July release

Venue.sh now frees memory as it processes each file, so the dashboard stays responsive while it prepares repository content for Assistant search and summaries.

• Fewer interruptions - Large repository ingestions are less likely to stop under resource pressure.
• No workflow changes - Your existing ingestion, search, and summary workflows work as before.