Skip to Content
GuidesGuides

Guides

Practical walkthroughs of every part of Shep.

Working with features

GuideRead it for
Your First FeatureA walkthrough from repo to merged PR.
Spec-Driven ModeOptional requirements and planning phases with approval gates.
CI Watch & Auto-FixWhat happens after the PR opens and CI runs.
Git checkpointsDiff or roll back any step an agent took.
PR review commentsLet an agent address review comments on its PR.
Multi-RepoRunning features across many repos from one place.

Interfaces

GuideRead it for
Web DashboardThe Control Center, Chat, Agent activity, the feature drawer and settings.
ChatAsking for work in a conversation, with inline specs, previews and plans.
Terminal consoleshep tui and shep ps — every feature from the keyboard.
CLI ReferenceEvery command, grouped by area.

Running at scale

GuideRead it for
Fleet managementParallel limits, the queue, triage, bulk approve and retry.
Supervisor & guardrailsPolicies and rules that approve or block gates for you.

Apps

GuideRead it for
Start an appCreate a new app from a description, with starters.
Dev serversHow Shep finds and runs your app’s dev server.

Agents & models

GuideRead it for
AgentsThe 12 agents Shep can drive, how to sign them in and switch.
Models & effortDefault model, reasoning effort, adaptive tiers and shep usage.
Agent harnessExperimental: Shep’s own query-aware harness and self-learning.

Integrations

GuideRead it for
MCPLet Claude Code or another MCP client drive Shep.
A2AThe A2A endpoint, the agent activity API and Herdr.

Configuration

GuideRead it for
ConfigurationWhere settings live, shep settings, environment variables.
Feature flagsWhat each flag turns on, and which are on by default.
Worktree provisioningCustom commands to create or prepare each worktree.
Usage metrics & privacyWhat Shep sends, what it never sends, and how to turn it off.

Building a larger pipeline — repos grouped into spaces, Linear or Jira sync, incidents, opportunities? See the Software factory section.

What’s next