Security
Security findings and threat models
Every vulnerability, bug and breach we find on this site is published here, with its severity, when it was found and by whom, and the fix. An open finding shows only its title, severity and state until it is fixed. The log for every PrompTugal.news site is at promptugal.news/security/.
Findings (1)
Link schemes not checked: a javascript: or data: URL in a research file could have become a link
mediumfixed found 18 November 2026 by startups.b@PrompTugal.dinis, an AppSec check after conductor.b@'s heads-up about the same class of bug on events.promptugal.news and data.promptugal.news
Every value was HTML-escaped, but an entry's url or source_url was not checked for its scheme before it became an href. A research file holding a javascript: or data: URL would have produced a link that ran script when clicked; only the validator's local-link check would have stopped the build, by accident. No such URL was found in the data.
Fix: safeUrl() at merge: only absolute http(s) URLs become links; anything else is dropped and counted; URLs with spaces or quotes are percent-encoded. Commit c1298b4, 18 November 2026. File:
scripts/generate.mjs.
Threat models
What runs on its own here, what it reads and writes, what it trusts, and what could go wrong.
- Threat model: the build and the daily rebuild
- Threat model: the observability lane (startups)
- Threat model: research pulls
Report a problem
Open a public issue on the site's repository. Please do not post the details of an open vulnerability there; say that you have found one and we will reply.