At a minimal level there should be 2 trivial checks any spreadsheet should be able to make on its formulas:
Ideally A well designed spreadsheet model should have a clear flow from input sheets to output sheets.
High Level Information
Auditing Maps
Provide visual documentation of a spreadsheets structure and contents. Also aid in finding your way around unfamiliar spreadsheets.
One way to do this is to produce a visual overview of all the cells in use, what's in them (formulas, numbers, errors, text, etc.).
Cell categories:
Example from [XLAUD].
Equivalence with cells in adjacent cells (and possible non-adjacent cells). Also consider top-originating areas as demonstracted in [MSA]. This could involve showing how formulas have been copied throughout a spreadsheet based on the equivalence classes.
Example from [MSA].
Highlight areas covered by range names. (Not currently supported by toolkit)
Display a statistical summary for each worksheet. A text listing of all the range names showing which cell they begin in and the areas they cover.
Cell Roots - Given that root cells are those which don't reference other cells themselves, collectivly they have the greatest effecton the rest of the spreadsheet. Small changes in their values can yield the largest changes elsewhere. As Alex suggested, a possibly good visualisation would show how many cells are dependent on a particular cell.
Circular References can currently be found using the toolkit. It's important to know which cells are involved, where they are located, if they cross worksheets (undesirable for a deliberate circular reference), and if iterative calculation has been enabled by the user.
Links