Team

Team Plugin - The Agentic Startup

Specialized agent library for Claude Code with consolidated, high-signal activities for software delivery.

The team plugin provides 8 specialized roles, 15 activity-based agents, and 14 reusable skills. This consolidation keeps specialist depth while reducing overlap and context bloat.

---

Installation

/plugin install team@the-startup

---

Agent Roles

The Chief

Complexity assessment and activity routing specialist

• the-chief

The Analyst (1 activity)

Product research and requirements specialist

Activity	Focus
research-product	Market analysis, requirement clarification, prioritization, stakeholder alignment

The Architect (4 activities)

System design and technical governance specialist

Activity	Focus
design-system	Scalable architecture, service boundaries, data + deployment strategy
review-security	Application security + dependency/supply-chain review
review-robustness	Complexity and concurrency risk review
review-compatibility	Breaking-change detection, migration safety, backwards compatibility

The Developer (2 activities)

Implementation and optimization specialist

Activity	Focus
build-feature	Feature implementation across UI, API, services, and data layers
optimize-performance	Bottleneck diagnosis and targeted optimization

The Tester (1 activity)

Functional and performance testing specialist

Activity	Focus
test-strategy	Risk-based test planning, quality coverage, load/stress validation

The Designer (3 activities)

User research and UX design specialist

Activity	Focus
research-user	User interviews, persona/journey insights, behavioral evidence
design-interaction	Information architecture, user flows, interaction patterns
design-visual	Design systems plus accessibility-by-default standards

The DevOps (2 activities)

Platform and operations specialist

Activity	Focus
build-platform	Containers + IaC + CI/CD as one delivery platform
monitor-production	Observability, SLI/SLOs, alerting, incident diagnostics

The Meta Agent

Agent design and generation specialist

• the-meta-agent

---

Skills System

Skills are referenced in agent YAML frontmatter:

---
name: agent-name
skills: project-discovery, pattern-detection, api-contract-design
---

Available Skills (14)

Category	Skill	Description
Cross-Cutting	project-discovery	Unified navigation, stack detection, and doc validation
	pattern-detection	Identify existing codebase patterns for consistency
	feature-prioritization	RICE/MoSCoW/Kano/value-effort prioritization
	requirements-elicitation	Requirement gathering, conflict resolution, acceptance criteria
Design	user-research	Research planning, evidence synthesis, personas, journey mapping
Development	api-contract-design	REST/GraphQL contract design and versioning
	architecture-selection	Pattern selection for monolith/microservices/serverless/event-driven
	domain-modeling	Domain + data modeling, invariants, schema evolution
	technical-writing	ADRs, system docs, runbooks, API docs
	testing	Layered testing strategy and execution patterns
Infrastructure	platform-operations	Pipeline design + observability + release reliability controls
Quality	code-quality-review	Holistic review including security/perf/a11y/error handling
	performance-analysis	Profiling and bottleneck identification
	security-assessment	Threat modeling and vulnerability review

---

Architecture Principles

1. Agents define who/what/when.
2. Skills provide reusable how.
3. Consolidation favors clearer boundaries and lower context overhead.
4. Deliverables stay concrete and execution-oriented.