Agents & subagents
PHRONEN can hand parts of a job to subagents that work in parallel, each with its own instructions and tools.
A subagent runs through the Task tool: PHRONEN gives it a description and a prompt, it works on its own and reports back. In the browser its steps appear nested under the Task card.
| Built in | What it is for |
|---|---|
| general-purpose | any multi-step job — research, coding, file work; independent parts of a task in parallel |
| explore | fast, read-only exploration: finds files, symbols and usages without changing anything |
| reviewer | reviews code changes for bugs, security problems and missing tests; runs the tests; read-only |
| web-tester | tests a running web app in a real browser: navigates, clicks through flows, checks console errors |
Your own
.phronen/agents/api-tester.mdmarkdown
---
name: api-tester
description: Exercises the HTTP API end to end and reports failures.
tools: Bash, Read, Grep, WebFetch
effort: fast
---
You are an API tester. Start the server if it isn't running, call every
endpoint listed in docs/API.md, and report each failure with the request,
the response, and the line of code responsible.- Front matter: name, description (how PHRONEN decides when to use it), optional tools (leave out for every tool) and effort; then the instructions.
- Project agents live in .phronen/agents/, yours in ~/.phronen/agents/.
- /agents lists them; the browser shows them under Agents.
Was this page helpful?