Author(s): Udaykiran Estari

Originally published on Towards AI.

The Problem Isn’t AI Skill — It’s AI Compliance

Claude already knows it should write a spec before touching code. It will tell you exactly that if you ask. But the moment you say “go build it,” it skips straight to implementation — every time. That gap between comprehension and compulsion is the exact problem Superpowers was designed to close, and it’s a problem that no amount of better prompting will fix.

After introducing the compliance gap, the article explains that Superpowers is less a productivity booster and more a process enforcement mechanism: it uses composable SKILL.md files that auto-trigger to phase-gate work through a strict 7-step workflow (Brainstorm → Spec → Plan → TDD → Subagent-Driven Execution → Two-Stage Review → Finalize), so the model can’t jump straight from a task to code without spec, planning, and tests. It details what each phase prevents (from missing architectural/data-model decisions to skipping TDD and enabling context bleed), the measurable quality benefits, and the real trade-offs—time overhead, higher token consumption, dependence on correct requirements, and the limitation that enforcement is in-context rather than a hard runtime constraint. It then provides a decision framework comparing Superpowers with gstack and GSD based on failure modes (regressions/quality, context bleed, or architecture/design bottlenecks), plus guidance for rollout (start with one complex project, write first custom SKILL.md files, and adopt whole-team rather than partially). Finally, it concludes with a “ship it or skip it” verdict: Superpowers is worth adopting when production-grade, multi-file work makes regressions and unintended behavior the primary failure mode, but should be skipped for quick fixes or teams that can’t maintain uniform workflow adoption.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI