Update
With newer biblatex versions we can simply redefine the macro \abx@missing@entry to change what is shown when a citation is not found.
\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage[backend=biber, style=numeric]{biblatex}
\usepackage{xcolor}
\addbibresource{biblatex-examples.bib}
\makeatletter
\def\abx@missing@entry#1{%
{\bfseries\color{red} undefined Label: #1}}
\makeatother
\begin{document}
\cite{cicero}
\cite{lorem}
\end{document}
Old answer
With biblatex, we have to modify \blx@citation@entry, a command called for each citation item:
\makeatletter
\def\blx@citation@entry#1#2{%
\blx@bibreq{#1}%
\ifinlist{#1}{\blx@cites}
{}
{\listgadd{\blx@cites}{#1}%
\blx@auxwrite\@mainaux{}{\string\abx@aux@cite{#1}}}%
\ifinlistcs{#1}{blx@segm@\the\c@refsection @\the\c@refsegment}
{}
{\listcsgadd{blx@segm@\the\c@refsection @\the\c@refsegment}{#1}}%
\blx@ifdata{#1}
{}
{\ifcsdef{blx@miss@\the\c@refsection}
{\ifinlistcs{#1}{blx@miss@\the\c@refsection}
{{\bfseries\color{red} undefined Label:} }
{\blx@logreq@active{#2{#1}}}}
{\blx@logreq@active{#2{#1}}}}}
\makeatother
To get rid of a spurious space, we can also redefine (this has to be placed within a \makeatletter ... \makeatother group)
\def\blx@citeadd#1{%
\ifcsdef{blx@keyalias@\the\c@refsection @#1}
{\edef\blx@realkey{\csuse{blx@keyalias@\the\c@refsection @#1}}}
{\def\blx@realkey{#1}}%
\expandafter\blx@citation\expandafter{\blx@realkey}\blx@msg@cundefon
\expandafter\blx@ifdata\expandafter{\blx@realkey}
{\advance\blx@tempcnta\@ne
\listeadd\blx@tempa{\blx@realkey}}
{\ifnum\blx@tempcntb>\z@\multicitedelim\fi
\expandafter\abx@missing\expandafter{\blx@realkey}%
\advance\blx@tempcntb\@ne}}
Where we added a % at the end of the second-to-last line.
MWE
\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage[backend=biber, style=numeric]{biblatex}
\usepackage{xcolor}
\addbibresource{biblatex-examples.bib}
\makeatletter
\def\blx@citation@entry#1#2{%
\blx@bibreq{#1}%
\ifinlist{#1}{\blx@cites}
{}
{\listgadd{\blx@cites}{#1}%
\blx@auxwrite@mainaux{}{\string\abx@aux@cite{#1}}}%
\ifinlistcs{#1}{blx@segm@\the\c@refsection @\the\c@refsegment}
{}
{\listcsgadd{blx@segm@\the\c@refsection @\the\c@refsegment}{#1}}%
\blx@ifdata{#1}
{}
{\ifcsdef{blx@miss@\the\c@refsection}
{\ifinlistcs{#1}{blx@miss@\the\c@refsection}
{{\bfseries\color{red} undefined Label:} }
{\blx@logreq@active{#2{#1}}}}
{\blx@logreq@active{#2{#1}}}}}
\def\blx@citeadd#1{%
\ifcsdef{blx@keyalias@\the\c@refsection @#1}
{\edef\blx@realkey{\csuse{blx@keyalias@\the\c@refsection @#1}}}
{\def\blx@realkey{#1}}%
\expandafter\blx@citation\expandafter{\blx@realkey}\blx@msg@cundefon
\expandafter\blx@ifdata\expandafter{\blx@realkey}
{\advance\blx@tempcnta@ne
\listeadd\blx@tempa{\blx@realkey}}
{\ifnum\blx@tempcntb>\z@\multicitedelim\fi
\expandafter\abx@missing\expandafter{\blx@realkey}%
\advance\blx@tempcntb@ne}}
\makeatother
\begin{document}
\cite{cicero}
\cite{lorem}
\end{document}

grep -i undefined <filename>.logwill give you a list of lines likeLaTeX Warning: Citation 'bibkey' on page 134 undefined on input line 234.With similar things for unresolved cross references etc. With a littlesed,sortanduniqyou can parse this further if you wish to e.g. give you a simple list of all undefined bibkeys, or a list of all unresolved references to figures. Or whatever. That is, you have *your computer* filter the bits you are interested in so that *you* don't have to. For long documents, obviously - not for a couple of pages. – cfr Jan 25 '15 at 03:47cleverefandbiblatexare two completely unrelated (safe for the goal, of course); so you might benefit from splitting this question in two. – moewe Jan 25 '15 at 08:16biblatexandbiberto work, so feel free to tell me what I should do withbiblatex. My idea was to use\@setcitein analogy with the other answers, but that doesn't seem to do it. – jdc Jan 27 '15 at 20:50??are easier to find. I wouldn't call thisvisibilitybut, hey, maybe your eyes work differently from mine. (Actually, this is overwhelmingly likely since my eyes do not work in the normal human fashion.) But if I'm searching rather than visually scanning, I wouldn't search for??or bibkeys. I wouldn't search the output at all. Plain text is so much easier to search and parse, I don't see the point in searching the DVI/PDF/PS. – cfr Sep 24 '15 at 21:48