Agent Foreman Plugins
Run long, multi-feature agent builds with external memory, spec-driven features, and clean handoffs between foreman, PM, UX, tech, and QA roles.
Overview
Agent-foreman-plugins is a plugin marketplace for the Build phase that provides a Long Task Harness with external memory, feature-driven skills, and multi-agent handoffs for sustained Claude Code projects.
What is this marketplace?
- Long Task Harness with external memory and feature-driven workflow (v0.2.1).
- Seven role agents: foreman, pm, ux, tech, qa, breakdown-writer, implementer.
- Five skills: project-analyze, init-harness, foreman-spec, feature-next, feature-run.
- Designed for clean agent handoffs on multi-session Claude Code work.
- MIT-licensed plugin from Lukin with repository at github.com/mylukin/agent-foreman.
- 1 plugin in marketplace
- 7 agent role definitions
- 5 workflow skills
Community signal: 248 GitHub stars.
What problem does it solve?
Multi-day agent builds lose context and overlap work because nothing externalizes specs, features, and QA handoffs.
Who is it for?
Solo builders running long Claude Code roadmaps who need PM-style specs, queued features, and separated implementer versus QA agents.
Skip if: One-off scripts, tiny fixes, or teams that already enforce full human project management outside the agent.
What do I get? / Deliverables
You run feature-next and feature-run cycles with persistent harness memory and role agents that hand off cleanly between sessions.
- Initialized harness with init-harness
- Foreman spec and prioritized feature queue
- Completed feature-run cycles with QA and implementer handoff records
Plugins in this marketplace
1 plugin — install individually after you add the marketplace.
Recommended Marketplaces
Journey fit
Canonical shelf is build/pm because the harness starts with project analysis, specs, and feature queues that govern how implementation agents work. PM subphase matches foreman-spec, feature-next, and feature-run skills that sequence work like a solo founder’s product lead.
How it compares
Long-task orchestration marketplace with seven agents, not a single coding skill or deploy MCP.
Common Questions / FAQ
Who is agent-foreman for?
Indie and solo developers using Claude Code on multi-feature products who need durable memory and structured agent roles.
When should I use agent-foreman?
Use it at the start of a sustained build when tasks span many sessions and you want spec-driven feature queues plus handoffs.
How do I add agent-foreman to my agent?
Add the Agent Foreman Plugins marketplace in Claude Code, install the agent-foreman plugin, then run init-harness and project-analyze per bundled skills.