| name | critique |
| disable-model-invocation | true |
| description | Systematically find problems in a design before building it. Four lenses:
unnecessary complexity, missing fundamentals, feasibility gaps, and scope
mismatch. Use after creating an implementation plan or when a design feels
wrong. Invoked via /critique.
|
Critique
Apply four lenses to the design. Each lens asks different questions.
Lens 1: Unnecessary Complexity
- What could be removed without losing the core value?
- Which components exist for hypothetical future needs?
- Where are abstractions hiding simple operations?
- What would the simplest version look like?
Lens 2: Missing Fundamentals
- What happens on the first error?
- What happens at 10x the expected load?
- How does this fail when a dependency is down?
- What's the recovery path from data corruption?
- Where are the implicit assumptions?
Lens 3: Feasibility Gaps
- Which components have we never built before?
- Where are we estimating based on hope rather than measurement?
- What are the integration boundaries we haven't tested?
- Which third-party dependencies are we trusting blindly?
Lens 4: Scope Mismatch
- Does the implementation match what was actually asked for?
- Are we building Level 2 (polish) before Level 0 (platform)?
- What's the gap between the demo and production?
- Which features are solving the wrong problem?
Output Format
For each problem found:
- Which lens caught it
- The specific problem (one sentence)
- The risk if unaddressed
- A concrete fix (not "think about it more")
Number problems sequentially. The list IS the deliverable.
