# Review Team > Compose and orchestrate dynamic review teams using Claude Code agent Use when this capability is needed. - Skill: `tomevault-io/review-team-2` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/review-team-2` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/review-team-2/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: AI & ML - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/review-team-2 --- # Review Team — Index Compose a multi-agent review team: pick agents, create the team, coordinate, synthesize. Each agent's full brief — scope, focus areas, how it reports — lives in its definition file at `agents/.md`. Spawn teammates with task-specific context from there. Agents are generic and stack-agnostic; every one inherits the session's tools, so any agent can both review and do the work in its domain. ## Agents | Agent | Specialty | |-------|-----------| | `security-analyst` | App, infra & AI/data security; vulns, secrets, auth, supply chain | | `software-engineer` | Code quality, correctness, idioms, tests — any language | | `platform-engineer` | CI/CD, IaC, containers, deployment, observability, DX | | `data-engineer` | Pipelines, schema & data modeling, ETL, data quality, migrations | | `ai-specialist` | LLM/agent systems, prompts, evals, retrieval, guardrails | | `qa-technician` | Test strategy, coverage, flakiness, regression — any stack | | `product-manager` | Requirements, scope, user value, acceptance criteria, trade-offs | | `frontend-designer` | UI/UX, interaction design, responsiveness, accessibility | | `legal-expert` | OSS licensing, ToS, data privacy/compliance, attribution | | `technical-writer` | Docs accuracy, READMEs, API docs, ADRs, onboarding | | `devils-advocate` | Challenges the direction itself — assumptions, consensus, the rejected option (meta-review, not domain review) | ## Recipes Starting points — drop agents that don't apply, add any the work touches. | Recipe | Agents | When | |--------|--------|------| | Quick | `software-engineer` + `security-analyst` | Routine PR, small change | | Feature | `product-manager` + `software-engineer` + `qa-technician` | New feature before merge | | Frontend | `frontend-designer` + `software-engineer` + `qa-technician` | UI work, design + a11y review | | Data / AI | `data-engineer` + `ai-specialist` + `security-analyst` | Data pipeline or LLM feature | | Platform | `platform-engineer` + `security-analyst` + `software-engineer` | CI/CD, IaC, or deployment change | | Release | `security-analyst` + `legal-expert` + `technical-writer` | Pre-release licensing/security/docs audit | | Comprehensive | most of the roster | Major service / architectural review | `devils-advocate` is not a domain reviewer — add it to any recipe for a high-stakes, hard-to-reverse, or gate decision (ship / merge / architecture sign-off). Run it **after** the domain reviewers report and their findings are consolidated, not in parallel — it needs a settled direction to attack. Skip it on routine PRs. ## Running a team Requires `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1`. Create the team with one task per review dimension; spawn each teammate with task-specific context (they don't inherit conversation history) and ask for severity-rated findings with `file:line` references. The lead coordinates and synthesizes findings — it does not implement. **Synthesize** the teammates' reports into one severity-ordered summary: dedupe overlapping findings (note which agent raised each), surface Critical/High first, and flag any disagreements between agents rather than silently resolving them. **When not to use a team:** teams cost several times more than a single agent. For small, sequential, or token-sensitive work — or a single-dimension check one agent finishes in under ~5 minutes — use one agent or a plain subagent instead. --- > Source: [jedwards1230/claude-plugins](https://github.com/jedwards1230/claude-plugins) — distributed by [TomeVault](https://tomevault.io).