superpowers-patterns

To enforce quality controls in AI task execution, apply configurable iron laws and hard gates that block progress until conditions are met. This prevents skipped steps and ensures verifiable results for complex workflows.

Hard gates and iron laws are non-negotiable quality controls for multi-step task integrity. Iron laws are absolute rules that cannot be bypassed, while hard gates act as checkpoints blocking progress until specific conditions are verified.

To stop AI from rationalizing shortcuts, use rationalization tables and anti-slop rules that preempt self-justification with reality checks. These explicitly define patterns to avoid vague claims and halt progress when quality is at risk.

Yes, you can use checklist enforcement for complex research and planning workflows. It mandates completion and verification of steps, applying red-flag checks to guard against shortcuts where step integrity matters.

Yes, enforcing anti-slop rules works without external dependencies. The enforcement mechanism relies entirely on internal configurable rules, rationalization tables, and checklist verification to maintain step integrity.

You should not use iron-law quality enforcement for simple, single-step tasks. It is designed for complex, multi-step coding, research, and planning workflows where hard gates and step integrity are necessary to prevent shortcuts.