Orchestrated planning, security auditing, and code quality skills for Claude Code
Agent
Orchestrates parallel security audit across 20+ specialized audit skills against Solidity codebases. Creates teams, spawns wave-based teammates that invoke audit skills via Skill tool, synthesizes deduplicated findings report. Use for comprehensive smart contract security analysis.
Agent
Fixes build/compilation errors with surgical precision across TypeScript, Rust, and Solidity. Minimal diffs only — gets builds green without architecture changes. Emergency unblocking when builds break. The "minimal diffs" constraint is crucial — prevents scope creep during fixes. Diagnostic categorization helps…
Agent
General code review: CLAUDE.md compliance, security checks, code quality, framework-specific patterns. Confidence-based filtering (>=80%). Every PR. Broadest coverage of any review agent. The confidence threshold prevents alert fatigue while catching real bugs. Multi-language, multi-framework.
Agent
Generates and updates documentation and codemaps from source code. Uses language-specific analysis tools (TypeScript LSP, cargo-modules, slither) for AST analysis, call chain tracing, dependency mapping. Keeping documentation synchronized with code. Prevents documentation rot. Output still needs human verification.
Agent
CI/CD pipelines, GitOps, container security, zero-downtime deployments, progressive delivery. Use PROACTIVELY for deployment automation, infrastructure config, or production readiness checks.
Agent
Iteratively refines UI through screenshot-analyze-improve cycles. One change per iteration, includes competitor research. Polishing working implementations that don't feel right. The one-change-per-cycle discipline prevents over-engineering. Anti-AI-slop guidance produces distinctive designs.
Agent
Draw.io diagram specialist — generates diagrams from codebase analysis, verifies diagram logic against source code, and fixes routing/layout issues. Used standalone for creation or by /orc:compound-diagram for audit and verification tasks.
Agent
Captures Figma specs + implementation screenshots, finds visual discrepancies, and fixes them with CSS/Tailwind changes. Any Figma-to-code workflow. Iterative — run repeatedly until pixel-perfect. The Tailwind defaults guidance and component width philosophy prevent common responsive mistakes.
Agent
Verifies frontend via browser automation — forms, interactions, visual regressions, responsive layouts. Also fixes issues found. General frontend QA when there's no Figma involved. Only non-Figma design agent. Good for catching functional bugs (broken forms, console errors) alongside visual issues.
Agent
Archaeological analysis of git history: traces code evolution, identifies contributors, understands why patterns exist. Understanding historical context of complex or puzzling code. Niche but valuable when you need to know why code looks the way it does. Requires clear motivation — not a proactive agent.
Agent
Monitors phase plans for ADR compliance and flags architectural drift asynchronously. Non-blocking — runs in parallel with planner-checker convergence loop.
Agent
Explores codebase and writes structured analysis documents. Spawned by /orc-map with a focus area (tech, arch, quality, concerns). Writes documents directly to reduce orchestrator context load.
Agent
Plan-first task execution. For each task, spawns a planner then specialist subagents for implementation. Handles git commits, verification, and team coordination. Persistent teammate in implementation teams.
Agent
Plan-first implementation verification. Spawns a verification planner, then executes each verification step via specialist subagents. Persistent teammate in implementation teams.
Agent
Researches how to implement a phase before planning. Produces RESEARCH.md consumed by orc-swarm. Spawned by /orc-phase-research orchestrator.
Agent
Verifies phase plans meet quality gates using 9 verification dimensions. Persistent teammate in plan-phase teams.
Agent
Creates executable phase plans with goal-backward verification. Persistent teammate in plan-phase teams.
Agent
Researches domain ecosystem before roadmap creation. Produces files in research/ consumed during roadmap creation. Spawned by /orc-research orchestrator.
Agent
Synthesizes research outputs from parallel researcher agents into SUMMARY.md. Spawned by /orc-research after 4 researcher agents complete.
Agent
Identifies design patterns (Factory, Observer, Strategy), anti-patterns (God objects, circular deps), naming consistency, code duplication via jscpd. Periodic codebase audits for consistency. Nice-to-have rather than must-have — pattern detection is lower priority than correctness. May require jscpd setup for…
Agent
Application performance optimization — observability, profiling, load testing, caching, Core Web Vitals. Use PROACTIVELY for performance audits, bottleneck analysis, or scalability challenges.
Agent
Systematic analysis of repository structure, conventions, documentation, templates, and codebase patterns. Onboarding to a new codebase or understanding an unfamiliar project's conventions. More comprehensive than a quick code search but less actionable than learnings-researcher.
Agent
OWASP Top 10, hardcoded secrets, injection attacks, unsafe crypto, dependency audits, framework-specific patterns (Node.js, Rails, Rust, Solidity). Emergency response protocols. Any code handling user input, auth, payments, or sensitive data. Dependency scanning (npm audit, bundle audit, cargo audit, slither) catches…
Agent
Zero-tolerance error handling audit: empty catches, broad exceptions, stale closures, inadequate user feedback, missing logging. JS/TS try-catch, Rust Result/Option misuse, Solidity unchecked call returns. Prevents debugging nightmares from swallowed errors. The focus on user-facing error quality prevents "something…