Skip to content

Security: zaid-mian/vibe-coder

Security

SECURITY.md

Security policy

Scope

This repository contains no application code, no servers, and no user data. It is a set of Markdown instructions plus dependency-free Python scripts that read files from a local repository.

A vulnerability here means one of:

  • a script that can be made to leak file contents it should not read (for example, printing a secret's value from a scanned repository)
  • a script that can be made to write, delete, or modify files outside the directory it was pointed at
  • an instruction in a mode file that could cause an AI agent to perform a destructive or security-reducing action without human approval
  • a supply-chain issue: a dependency added where none is needed

What this project deliberately does not do

  • No network calls in any script. No telemetry, no analytics, no phone-home.
  • No third-party dependencies. Standard library only, so there is no dependency supply chain.
  • No secret values are ever printed by scan_project.py — it reports file path, line number, and the pattern name only. That behaviour is intentional and a regression in it is a security bug.
  • No credential rotation, IAM change, or infrastructure modification is ever performed or recommended as an automatic action.

Reporting

Report privately. Do not open a public issue for a leak or an unsafe instruction, since the report may describe an exploitable path.

Please include:

  • what you did, step by step
  • what you expected, and what happened instead
  • the affected file and line
  • the command output, with any real secret removed

You will get an acknowledgement, an assessment of severity, and a fix or a written explanation of why the behaviour is intended.

If you find a real secret in a repository you scanned

That is a finding about that repository, not this one. Do not paste the value anywhere — not in an issue here, not in a message. Report it to the maintainers of the affected repository through their private channel, and recommend rotation, because a committed credential must be considered compromised. See references/mode-security.md for the checklist this project uses.

Not in scope

  • The behaviour of the AI assistant you load this skill into.
  • The security of a project you scanned with these tools.
  • The correctness of your own planning documents.

There aren't any published security advisories