Gathers empirical evidence for a hypothesis: benchmarks, case studies, production reports, documentation, community feedback. Rates each piece by source reliability and contextual congruence. Example: "Found TechEmpower benchmarks showing 2x throughput — R:0.9, CL:0.7." Negative: does NOT make decisions or verify logic.
Generates one grounded hypothesis for a decision question from a specific angle. Studies the project first, then proposes a concrete option with claims and risks. Example: "Angle: performance → Hypothesis: Use ClickHouse because write throughput matters most here." Negative: does NOT verify or gather evidence — only proposes.
Logically verifies a hypothesis against constraints, consistency, and known failure modes. Checks internal logic only — no empirical evidence gathering. Outputs a verdict: PASS (L1), FAIL (Invalid), or PARTIAL (stays L0). Example: "H1 claims ClickHouse fits, but the project needs ACID transactions — FAIL." Negative: does NOT propose alternatives or gather benchmarks.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
English | Русский
A Claude Code plugin marketplace for AI agent teams, expert debates, deep planning, and interactive code audits
Install focused plugins that make Claude Code more predictable for delivery, decisions, and cleanup.
izteam is an independent plugin marketplace for Claude Code. Each plugin adds slash commands, agents, and ready-to-use workflows: from building features with an AI team to auditing outdated code.
graph TD
subgraph Install
A["1. Add marketplace<br/><code>/plugin marketplace add izzzzzi/izTeam</code>"]
B["2. Install plugins<br/><code>/plugin install team@izteam</code>"]
A --> B
end
B --> C{"Choose a workflow"}
subgraph "/build — Team"
D1["📋 Brief — interview to clarify scope"]
D2["📐 Plan — architect designs tasks"]
D3["👨💻 Code — parallel coders implement"]
D4["🔍 Review — security + logic + quality"]
D5["✅ Commit — approved changes merged"]
D1 --> D2 --> D3 --> D4 --> D5
end
subgraph "/think — Think"
E1["🔬 Decompose into aspects"]
E2["🧠 Parallel expert analysis"]
E3["📄 Unified design document"]
E1 --> E2 --> E3
end
subgraph "/arena — Arena"
F1["📚 Research context"]
F2["🎭 Experts debate"]
F3["🤝 Consensus reached"]
F1 --> F2 --> F3
end
subgraph "/audit — Audit"
G1["🔎 Scan codebase"]
G2["💬 Interactive review with user"]
G3["🧹 Cleanup confirmed items"]
G1 --> G2 --> G3
end
C --> D1
C --> E1
C --> F1
C --> G1
Example workflows:
# Build a feature with an AI team
/build "Add user settings page with profile editing"
# Think through architecture before coding
/think "Migrate from REST to GraphQL — trade-offs and plan"
# Get expert opinions via structured debate
/arena "Microservices vs monolith for our SaaS?"
# Find and clean up dead code
/audit
| Plugin | Version | Description | Command |
|---|---|---|---|
| 🤖 team | 0.3.7 | Build features with an AI agent team and built-in review gates. | /build |
| 🧠 think | 1.1.7 | Plan complex tasks before coding with structured expert analysis. | /think |
| 🎭 arena | 1.1.7 | Compare expert viewpoints and converge on a clear decision. | /arena |
| 🧹 audit | 0.1.10 | Find dead and outdated code with an interactive audit. | /audit |
/plugin marketplace add izzzzzi/izTeam
/plugin install team@izteam
/plugin install think@izteam
/plugin install arena@izteam
/plugin install audit@izteam
Plugins are loaded on startup, so restart after installation.
Build features with an AI agent team and built-in review gates.
Required:
CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1insettings.json
/plugin install team@izteam
Commands:
/build "Add user settings page"
/build docs/plan.md --coders=2
/build "Add notifications" --git-checkpoints
/brief "Notifications system"
/conventions
Natural language — also works:
"Build a settings page with profile editing"
"Implement notifications with email and push"
"Interview me before we start building"
"Extract project conventions and document standards"
Repo Map feature inspired by Aider — generates a ranked symbol map to give agents instant codebase context.
Plan complex tasks before coding with structured expert analysis.
/plugin install think@izteam
Commands:
npx claudepluginhub izzzzzi/izteam --plugin reasonFind dead and outdated code with an interactive audit.
Build features with an AI agent team and built-in review gates.
Compare expert viewpoints and converge on a clear decision.
Plan complex tasks before coding with structured expert analysis.
Collect and synthesize opinions from multiple AI Agents for Claude Code
Deep analysis mode with extended reasoning for complex problems
Verification-first engineering toolkit for Claude Code. 15 skills across a 5-phase spine (Investigate → Design → Implement → Verify → Ship), 8 specialist agents, an interactive setup wizard. Every skill has rationalizations + evidence requirements. Built for senior ICs and tech leads.
Multi-agent deliberation for AI coding assistants
Meta-cognition: refine input through brainstorming, refine output through challenge and condensed communication mode.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.