Remediation Fix-Impact Simulator

“If I fix these surfaces, what does my scorecard look like AFTER?”

The unified scorecard tells you what is broken; the per-surface remediation guidance tells you how to fix each finding. This closes the loop: pick the surfaces you commit to fixing and AgentSmack re-runs the same proven scorecard with those surfaces projected to a clean ceiling, then diffs current → projected — the overall delta, the band transition (not-ready → needs-hardening → production-ready), which blocked dimensions clear, and which labs you still have to fix.

Two properties make the projection trustworthy. Honest-empty — a surface you never ran can't be “fixed”, so checking it is a no-op (it never fabricates an improvement). Unforgeable — the projected scorecard runs through the unchanged scorer, so a remediated-but-still-low stack cannot be washed into a fake pass, and any blocker you don't fix still caps the verdict. It also computes the smallest set of fixes that clears your band — the scorecard-level sibling of the kill-chain min-cut.

Want the underlying verdict first? Compose the unified scorecard → Or see whether those findings chained into a real compromise? Reconstruct the attack path →