broneq

Generate comprehensive architecture documentation for complex code modules with Graphviz diagrams and examples. Use when user asks to "explain this code", "document the architecture", or wants to understand how a module/system works.

Audit a system prompt or skill file against agent-architecture best practices. Returns scored findings; rewrite is opt-in.

Create Architecture Decision Records (ADRs) based on MADR template. Use when the user asks to "create an ADR", "document a decision", "write an ADR", or provides decision context that needs to be formalized.

Lint a BDK skill or agent file for best practices and portability. Use when reviewing, creating, or editing a skill (skills/*.md) or agent (agents/*.md) to verify it meets BDK conventions. Trigger on phrases like "check this skill", "verify agent", "does this follow best practices", "lint skill", or when about to commit a new skill/agent.

Execute an implementation plan, then run tests and static analysis, fixing issues

Design pattern principles (GoF, anti-patterns, pattern documentation). Preloaded into agents that propose or review code structure; not user-facing.

Use when existing architecture documentation needs refreshing after code changes, when running periodic doc maintenance, or when user asks to "update docs", "refresh documentation", or "sync docs with code".

Initialize .bdk/settings.json for this project. Run once per project when BDK blocks session start with missing settings.

Tool-tier guidance for codebase exploration and architecture overview. Preloaded into agents that explore code structure; not user-facing.

Verify implementation plans before execution using a 4-agent pipeline. Use when you have a written plan (from /bdk:create-plan or manual) and want to check if it will actually work before writing code.

Tool-tier guidance for symbol search and call/reference tracing. Preloaded into agents that locate code; not user-facing.

Coordinator that executes a plan in parallel groups via background subagents. Spawns implementers, then dedicated test-runner / static-analyse subagents, then routes failures back to the original implementer (SendMessage) or a fresh fixer. Fully autonomous — no human-in-loop.