deep-plan
Use before touching code for any multi-step task — features, refactors, migrations, architectural decisions, bug fixes, or any work that needs to be broken into verifiable steps. Trigger phrases include "write a plan", "implementation plan", "plan this", "plan the work", "break this down", "how should I approach this", "make a plan", "design the approach", "before I code this", "I need a plan", "write up the plan", "spec to plan", "requirements to plan". Produces an ADR-backed plan with verification-backed acceptance criteria via a Planner → Architect → Critic consensus loop.
Repository SourceNeeds Review
team
Use when coordinating multiple agents on a staged pipeline — plan → PRD → exec → verify → fix — with independent critic/verifier gates and two-stage review on every source modification. Trigger phrases include "spawn a team", "team of agents", "coordinate agents", "staged pipeline", "multi-agent pipeline", "agent team", "run a team on this", "PRD-driven team", "team workflow", "agents working together", "assemble a team", "delegate to a team", "orchestrate agents", "pipeline of agents". File-based state, no external MCP dependencies, honest termination labels, Claude Code native team tools.
Repository SourceNeeds Review
autopilot
Use when running full-lifecycle autonomous execution from a vague idea to working verified code — idea to battle-tested design to consensus plan to executed code to audited defects to three independent judge verdicts to honest completion report. Trigger phrases include "autopilot", "build me end to end", "full lifecycle", "idea to working code", "auto-run this project", "run this autonomously", "just build it", "go from idea to code", "do everything", "autonomous execution", "end-to-end build", "build this for me", "make it real end to end", "full autonomous build". Iron-law phase gates between every stage; no coordinator self-approval; honest termination labels.
Repository SourceNeeds Review
deep-debug
Use when a bug, test failure, or unexpected behavior needs diagnosing — including production incidents, regressions, stack traces, mysterious failures, flaky tests, or any symptom needing root-cause analysis. Trigger phrases include "debug this", "why is this failing", "find the bug", "fix the bug", "root cause", "what's wrong with", "this is broken", "diagnose", "troubleshoot", "investigate this failure", "the test is failing", "this used to work", "why doesn't this work", "where's the bug". Adversarial hypothesis-driven debugging with parallel competing hypotheses across orthogonal dimensions, blind independent judging, discriminating probes that falsify leaders, TDD-gated fix loops, and mandatory architectural escalation after 3 failed attempts.
Repository SourceNeeds Review