Skip to content

Security: schmug/shipofclaudius

SECURITY.md

Security Policy

shipofclaudius is a collection of dynamic workflow scripts for the Claude Code Workflow tool, packaged as a Claude Code plugin. There is no running service — the deliverables are the .claude/workflows/*.js scripts, their wrapper skills, and the plugin manifests. This policy covers vulnerabilities in those artifacts.

Reporting a vulnerability

Please report privately — do not open a public issue. A public issue discloses the vulnerability before a fix is available.

Report through GitHub's private vulnerability reporting:

This opens a private advisory visible only to you and the maintainers. Please include:

  • which workflow / file / manifest is affected,
  • a description of the issue and its impact,
  • steps to reproduce (a minimal args payload or crafted input, where relevant).

In scope

  • The workflow scripts in .claude/workflows/ — especially prompt-injection weaknesses in the workflows that read attacker-writable text (issue/PR bodies, comments, reviews, diffs, external findings/SARIF/CVE/GHSA, Dependabot alerts). See the Security model in README.md for the intended three-part defense; a bypass of that defense is in scope.
  • The wrapper skills in skills/ and the plugin manifests in .claude-plugin/ — e.g. a script/skill mismatch or manifest issue that could cause the wrong code to run.

Out of scope

  • Vulnerabilities in Claude Code, the Workflow runtime, or any tool a workflow invokes (gh, git, third-party scanners) — report those to their respective projects.
  • Findings that require the operator to already have write access, or to run a workflow against input they control and trust.
  • The offline test simulators in tests/, which use only Node built-ins and run no untrusted input.

Response timeline

This is a volunteer-maintained project, so timelines are best-effort:

  • Acknowledgement: within ~7 days.
  • Assessment & fix plan: within ~30 days of acknowledgement.

We'll keep you updated through the advisory thread and credit you in the fix (unless you'd prefer to remain anonymous).

There aren't any published security advisories