Act 4 First Agent — Development agent for Claude Code
Read-only narrative steward for Act IV (slides 16-18) — the First Agent demo that proves resumable streams via the F5 refresh beat.
How to install Act 4 First Agent
Installs to ~/.claude/agents/vercel-labs-workflow-workshop-act-4-first-agent.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/vercel-labs/workflow-workshop/HEAD/.claude/agents/act-4-first-agent.md -o ~/.claude/agents/vercel-labs-workflow-workshop-act-4-first-agent.md Restart Claude Code, or start a new session, for it to be picked up.
What Act 4 First Agent does
name: act-4-first-agent description: Read-only narrative steward for Act IV (slides 16-18) — the First Agent demo that proves resumable streams via the F5 refresh beat. Foundation act for Acts V and VI. tools: Read, Grep, Glob
You are the narrative steward for **Act IV — First Agent** of the Workflow SDK GA workshop deck. This act proves that DurableAgent streams survive a browser refresh. It's the foundation the next two agent acts build on.
Your slides (3)
| # | Route | Beat | |-
Alternatives in Development
- Turbo Agent — Implements Turbo Drive, Turbo Frames, and Turbo Streams for fast, responsive Rails applications with minimal J 652 ★
- Helm Conductor — Use this agent to orchestrate engineering work cycles — decide what to work on next, sequence stories and feat 523 ★
- Dsp Agent — Implement audio processing and DSP algorithms for Stage 2. Use PROACTIVELY after foundation-shell-agent comple 209 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Act 5 Observer
Read-only narrative steward for Act V (slides 19-21) — the Observer agent demo that proves durable tool-call r
Act 2 Verbs
Read-only narrative steward for Act II (slides 6-14) of the Workflow SDK GA deck. Covers the three workflow sc
Demo Producer
Storyteller and producer for polished product demo videos. Turns your narrative into a scripted, narrated, mus
Litreview Updater
Use to incrementally refresh an EXISTING literature review (from a prior /litreview run) with newly published
Plan Polisher
Create or refresh an /aif-plan plan, critique it, and run one refinement round at most. The caller launches an
Docs Refresh
Related Skills
Agent Session State
Durable, runtime-agnostic agent session state for React: approvals that survive reload, resumable streams, sub
Al Foundation Docs
Build or refresh both AGENTS.md and ARCHITECTURE.md as a consistent baseline for correct agent behavior
Demo Cmd Askme
Interview me before you act, asking only the questions that would change the output