Isolated Worker — Testing & QA agent for Claude Code
Implements a small, well-specified change once the scope and design are already decided — a bounded fix, a contained feature, a targeted refactor, an added test.
How to install Isolated Worker
Installs to ~/.claude/agents/arcanedge-ai-coding-agent-playbook-claude-code-isolated-worker.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/ArcanEdge-AI/coding-agent-playbook-claude-code/HEAD/agents/isolated-worker.md -o ~/.claude/agents/arcanedge-ai-coding-agent-playbook-claude-code-isolated-worker.md Restart Claude Code, or start a new session, for it to be picked up.
What Isolated Worker does
name: isolated-worker description: Implements a small, well-specified change once the scope and design are already decided — a bounded fix, a contained feature, a targeted refactor, an added test. Use when you can state the intended end state in a sentence or two and name the files involved. Not for work where the design is still open, the requirements are ambiguous, or the blast radius is unknown. model: sonnet effort: high permissionMode: default tools: Read, Grep, Glob, Edit, Write, Bash
Alternatives in Testing & QA
- Gitnexus Test CI Verifier — GitNexus test and CI reviewer 45.8k ★
- Review Test Structure — Reviews test code for isolation, independence, readability, documentation quality, and proper setup/teardown 432 ★
- Test Guardian — Monitors test coverage gaps when testable code is added or modified 34 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Worker Sonnet
Tier-2 worker on Sonnet. Use for well-specified, bounded work - implementing a decided design, writing tests,
TDD Coder
Implements one concrete, already-scoped piece of MeowPay work test-first - writes the failing test from a tdd-
Fast Worker
Mechanical, low-risk bulk work, mechanical edits, inventory sweeps, repetitive migrations, generated-file main
Fast Coder
Fast implementer for narrow, well-specified tasks with crisp acceptance criteria, such as an isolated endpoint
Mid Task
Executes moderately complex tasks (focused multi-file edits, contained bug fixes, small features, test writing
Artifact Sonnet
Use proactively for any HTML artifact page, new or revised, once its content is decided: a report, a consultat
Related Skills
Agent Skills Discovery RFC
A mechanism for discovering Agent Skills using the .well-known URI path prefix as specified in RFC 8615 for di
Subtask.Prompt
Delegate a bounded child task to a specialist agent and get a structured result back — use for isolated, well-
Playbook
How to dispatch well — choosing flash vs pro, writing self-contained briefs, parallelism, verifying results, a