jfk13

Sounds like you're expecting the AI-based tools that are finding bugs to also provide fixes.

I've been dealing with a bunch of AI-generated (or at least -assisted) vulnerability reports lately. In many cases the reports include proposed patches to fix the issues.

It's been..... interesting. In many cases, the analysis provided in the report has been accurate and helpful. In some cases, the proposed patches have also been good, and we've accepted them with minimal or no changes.

In other cases, despite finding a valid issue, and even providing a good analysis of the problem, the AI tool's suggested patch has been, quite simply, wrong.

Careful review from somebody who really _understands_ the code -- and the wider context in which it is operating -- is still absolutely necessary. That's not always going to happen in an hour.

jfk13

3 mesi fadiscuss

Just to confirm, I'm seeing the same issue on other repos. Sometimes a subset of the expected PRs show up, other times none at all.

jfk13

7 anni fadiscuss

Oh, and bots aiming to take advantage of the above-mentioned people.

jfk13

7 anni fadiscuss

If it's a simple solution, why hasn't someone built it?

I suspect that any such product that starts to get a significant user base is quickly overrun by people looking for cheap hookups, on the one hand, or "sugar" relationships (at best) on the other.

jfk13

7 anni fadiscuss

If the strcp() macro is used with a function as the first argument (or an expression that has side-effects), it's not obvious at the call site (or, probably, intended by the programmer) that the argument will be evaluated twice.

E.g. suppose we write something like

    char* buf = strcp(((char*)malloc(4)), "foobar", 4);

expecting this to copy the beginning of the string "foobar" into a newly-allocated 4-byte buffer. Oops... this will actually call malloc twice (leaking the first buffer), and it won't have written the intended '\0' into the buffer that actually ends up getting used, so all bets are off...

If strcp were a function, its first argument would be evaluated just once, and it would work as intended.

jfk13

7 anni fadiscuss

This is pretty appalling, really:

"Overall, of the 83 firms known to have held data... 24% supplied personal information without verifying the requester's identity."

Want someone else's personal data? No need to "hack into" any systems; just ask for it!

Submissions

Black Hat: GDPR privacy law exploited to reveal personal data

comments