Text Marks the Spot is an NVDA add-on, pure Python, no native libraries. It reads page structure from NVDA's accessibility tree and moves the browse-mode cursor. It does not make network calls. It does not collect telemetry. The only persistent state is a per-site exclusion list saved in NVDA's normal config directory.
Even so, security issues are possible. Things worth reporting:
- vulnerabilities in release delivery (a tampered
.nvda-addonpackage, for example) - unsafe handling of URLs, hostnames, or DOM-derived strings inside the add-on
- dependency vulnerabilities (the add-on itself ships no third-party Python packages, but the build tooling does)
- anything that could cause unintended code execution, privilege misuse, or unsafe trust decisions inside NVDA
Only the latest released version gets security fixes. Older releases are not supported.
Do not open a public GitHub issue for a suspected security problem. Report it privately by email instead:
Helpful to include:
- a short summary of the issue
- affected version
- steps to reproduce, if known
- impact
- any suggested mitigation
If you are not sure whether something is a security issue or a regular bug, err on the side of reporting privately first.
What I will try to do:
- acknowledge the report within 7 days
- decide whether it is in scope
- coordinate a fix or mitigation if it is confirmed
This is a small project. Response times will vary. Good-faith reports are appreciated.
Please give me time to ship a fix before sharing details publicly.