/team-review — Parallel Full-Team Audit

Run a full audit of your existing codebase with 4 specialist agents firing simultaneously. What would take an hour sequentially takes 15 minutes in parallel.

Usage

/team-review                    # full audit — all 4 agents on entire codebase
/team-review src/auth/          # focus all agents on a specific directory
/team-review --changed          # audit only files changed since last commit
/team-review --pre-release      # adds release-captain for release readiness check

What Gets Launched in Parallel

| Agent | Focus | |---|---| | code-reviewer | Quality, patterns, dead code, naming, complexity, maintainability | | security-reviewer | OWASP Top 10, secrets, injection, auth gaps, CVEs in deps | | performance-optimizer | N+1 queries, unbounded loops, missing indexes, memory leaks, bundle size | | api-guardian | Breaking changes, missing validation, versioning gaps, contract drift |

With --pre-release, also adds:

• release-captain — release readiness: changelog, semver, rollback plan, smoke test coverage

How It Works

• Identify the target scope (whole repo, directory, or changed files)
• Spawn all 4 agents simultaneously with run_in_background: true
• Each agent scans its assigned concern independently
• Wait for all to complete
• Aggregate into a structured Team Review Report:

## TEAM REVIEW REPORT
Code Quality (code-reviewer)
CRITICAL: [list]
HIGH:     [list]
MEDIUM:   [list]
Security (security-reviewer)
CRITICAL: [list]
HIGH:     [list]
Performance (performance-optimizer)
HIGH:     [list]
MEDIUM:   [list]
API Contracts (api-guardian)
BREAKING: [list]
WARN:     [list]

TOTAL: X critical, Y high, Z medium
RECOMMENDATION: [ship / fix critical first / block]

• After report: "Which finding do you want to fix first?"

Severity Levels

| Level | Meaning | Action | |---|---|---| | CRITICAL | Security vulnerability, data loss, production breakage | Block — fix before merge | | HIGH | Bug, broken contract, significant perf regression | Fix before shipping | | MEDIUM | Maintainability debt, minor perf, code smell | Fix when convenient | | LOW | Style, suggestion | Optional |

When to Use

• Before any major PR merge
• Pre-release audit
• Onboarding to an unfamiliar codebase (run /team-review to get the full picture fast)
• After a large refactor
• Monthly health check on long-running projects

Time and Token Efficiency

Sequential audit (4 agents one by one): ~40–60 minutes, full context per agent. /team-review parallel: ~10–15 minutes, narrow focused context per agent. System prompt cached across all agents → each pays 10% on that prefix.

Arguments

$ARGUMENTS

Execute Now

Invoking this command is confirmation to proceed. Do not show this documentation to the user — execute immediately:

• Identify the target scope from $ARGUMENTS (directory, --changed, --pre-release). Default: entire current project.
• Spawn all agents simultaneously using run_in_background: true:

• Wait for all agents to complete.
• Aggregate results into the Team Review Report format above.
• End with: "Which finding do you want to fix first?"