Every Claude Code subagent — project, plugin, and global — in one sidebar, each tagged with its scope and the model it runs on. No hunting through .claude/agents/ to remember what you built.
Your subagents are files too. See them all at once.
Claude Code loads agents from your project, your installed plugins, and your global config. The extension reads every scope and puts them in one browsable list.
Every scope in one list
Project agents in .claude/agents/, agents shipped by plugins, and global agents in ~/.claude/ — normally scattered across directories. Here they sit together, each with a badge for where it comes from.
Project, plugin, and global agents unified
Scope badge on every agent
Jump straight to the source file
Model badges, no file-opening
Each agent shows the model it runs on, parsed from its frontmatter. Tell an Opus reviewer from a Haiku scout at a glance — before you delegate the wrong task to the wrong model.
Opus / Sonnet / Haiku badge per agent
Inherited-model agents flagged clearly
Read-only: never rewrites your definitions
FAQ:
Agent questions, answered.
Where does it find my agents?
Every scope Claude Code uses: project agents in .claude/agents/, plugin agents from installed plugins, and global agents in ~/.claude/agents/. All in one list, each badged by scope.
Can I see which model an agent uses?
Yes. Each agent shows a model badge from its frontmatter — Opus, Sonnet, Haiku, or inherited — without opening the file.
Does this change how agents run?
No. The extension only reads and displays your agent files. They still run through Claude Code exactly as before.
Is it 100% local?
Yes. It reads your local ~/.claude/ and project .claude/ and renders under a strict CSP. No network, no telemetry, no stored credentials.
Stop grepping .claude/agents/. Just look.
Free, open source, reads only your local ~/.claude/.