Skip to content

Security: Kyaa-A/doesntbreak

Security

SECURITY.md

Security Policy

doesntbreak is an agent skill and plugin. In addition to instructions, it ships a local browser-based responsive checker and a startup update checker. The browser checker executes against user-selected local pages. The update checker makes a best-effort HTTPS request to this repository's raw GitHub manifest, writes a version/timestamp cache only in the host-provided plugin data directory, and prints manual update commands. It does not download releases, install updates, or modify the plugin directory.

Reporting a vulnerability

Please report security concerns privately to asnaripacalna@gmail.com rather than opening a public issue. Include:

  • a description of the issue and its impact,
  • the file and lines involved, and
  • steps to reproduce, if applicable.

Expect an initial response within a few days. Please allow a reasonable window to address the issue before any public disclosure.

In scope

  • Content in SKILL.md or references/patterns.md that could induce an agent to produce insecure code (for example, a CSS/HTML snippet carrying an injection or data-exfiltration vector).
  • Misleading or hidden instructions that could cause unsafe behavior in a tool-using agent.
  • Command injection, unsafe path handling, unintended file access, or browser isolation failures in src/responsive-check.ts, its generated scripts/responsive-check.js, and its launch commands.
  • Update-checker network, manifest-validation, cache, hook-command, and terminal-output vulnerabilities in src/check-update.ts, its generated scripts/check-update.js, and hooks/.

Out of scope

  • Vulnerabilities in Claude, Claude Code, or any host application that loads this skill: report those to their respective vendors.
  • General responsive-design disagreements; those are regular issues or PRs.

Supported versions

This project tracks a single line of development on main. Fixes land on main; there are no separately maintained release branches.

There aren't any published security advisories