Skip to content

Security: aybvip/Pi-Visual-Builder

Security

SECURITY.md

Security Policy

There is no executable code in this repository yet

This repository is a design specification and a set of static HTML mockups. It contains no application code, no runtime, no dependencies and no deployed service, so there is no attack surface to report on today.

If a genuine security issue is found in something that does ship here (for example the static mockup pages), open a GitHub Issue describing it.

When an implementation exists

Once code is added to this repository, report vulnerabilities privately using the Report a vulnerability button on the GitHub Security tab. Do not open a public Issue for a vulnerability.

Until that button is configured, contact the maintainer through the contact listed on https://github.com/aybvip.

Proposed security considerations

The properties below describe the intended design in this specification. None of them is implemented or verified yet, and none should be relied upon.

Plugin security (proposed)

  • Plugins are sandboxed — cannot access other plugin state
  • External network requests require explicit permission in the manifest
  • No direct DOM manipulation (use framework APIs only)
  • Obfuscated code is rejected during review

Data security (proposed)

  • Privacy rules enforced at runtime
  • PiDCTP on-chain state treated as immutable — the visual builder cannot alter it
  • API keys stored encrypted and never exposed to client-side code

AI agent security (proposed)

  • AI actions require user confirmation before execution
  • AI cannot modify privacy rules without explicit approval
  • AI-generated code is reviewed before deployment
  • Rate limiting prevents abuse

Scope

In scope

  • Anything that ships from this repository's own source tree

Out of scope

  • Pi Network core protocol — report to the Pi team
  • Pi Browser — report to the Pi team
  • Pi Network Platform API — report upstream
  • Third-party services referenced in the specification — report upstream

There aren't any published security advisories