Skip to content

Security: GrubbyLee/skill-manager

Security

SECURITY.md

Security Policy

skill-manager inspects local Claude Code / Codex skill and MCP metadata. Most commands are read-only for AIDE data, but security reports are still welcome.

Supported Versions

The main branch and latest GitHub Release are supported.

Reporting a Vulnerability

Please open a private security advisory or contact the maintainer through GitHub:

https://github.com/GrubbyLee/skill-manager/security/advisories/new

If the issue is not sensitive, you can open a regular issue:

https://github.com/GrubbyLee/skill-manager/issues

Sensitive Data

Do not include API keys, passwords, cookies, tokens, private keys, full session logs, or private filesystem paths in public issues or discussions.

For MCP-related reports, describe the transport and command shape, but do not paste environment variable values.

Project Boundaries

  • skm scan, skm status, skm risks, skm report, skm graph, skm list, skm search, skm recommend, skm ask, skm dupes, and skm audit are designed to be read-only for Claude Code / Codex data.
  • sessions --clean, disable, and enable are the only file-modifying command families.
  • File-modifying commands must keep confirmation, backup, and safety-window safeguards.

There aren't any published security advisories