select-loop-strategylisted

# Select Loop Strategy: Orchestration Pattern Decision Tree Provides a deterministic decision framework to help agents and developers select the right execution topology for any given software engineering, research, or system evolution task. --- ## The Master Decision Tree Evaluate your task against the following gates in order: ``` [Incoming Task / Trigger] │ ▼ 1. Does the task require strict human approval gates, formal state tracking, transactional worktree isolation, or automatic rollbacks on test failure? ├─ YES ──▶ Pattern 7: graph-execution (Deterministic State Machine) └─ NO ──▶ continue │ ▼ 2. Can the work be partitioned into 10+ independent, non-overlapping items that execute simultaneously with zero shared state? ├─ YES ──▶ Pattern 4: agent-swarm (Parallel Fan-Out) └─ NO ──▶ continue │ ▼ 3. Is the primary requirement adversarial critique, security analysis, or multi-perspective red-teaming until an explicit "Approved" verdict? ├─ YES ──▶ Pattern 2: red-team-review (Generator / Critic Feedback) └─ NO ──▶ continue │ ▼ 4. Does the task involve unguided friction discovery, automated hypothesis testing, and headless benchmark evaluation over long horizons? ├─ YES ──▶ Pattern 5: triple-loop-learning (Meta-Learning System) └─ NO ──▶ continue │ ▼ 5. Does the task require separating strategy/git management (Outer Loop) from tactical coding/test execution (Inner Loop)? ├─ YES ──▶