Skip to content

Security: wyeworks/accountable-review

Security

SECURITY.md

Security Policy

Supported Versions

accountable-review is a Claude Code plugin distributed as a single rolling release, pinned by the version field in .claude-plugin/plugin.json. There are no long-term maintenance branches, so security fixes are only made against the latest release.

Version Supported
Latest release ✅
Older releases ❌

If you are running a pinned release (a marketplace entry pinned by ref or sha, or an older git tag), please update to the latest tag before reporting an issue, unless the report is about the pin mechanism itself.

Scope

This repository has no application code and does not process user data at runtime. The parts worth a security report are:

  • The generated GitHub Actions workflow (skills/setup-ci/templates/workflow.yml and the scripts under skills/setup-ci/scripts/ that render and install it) — its triggers, guards, permissions, and the one GITHUB_TOKEN usage it contains.
  • The shell scripts under scripts/, ci/, and skills/*/scripts/ that generate or validate a Review Map.
  • The prose instructions (SKILL.md, references/*.md, agents/claim-falsifier.md) — for example, a wording that would cause the skill to execute untrusted code from a pull request, exfiltrate repository contents, or post somewhere it should not.

General Claude Code platform vulnerabilities (unrelated to this plugin) should be reported to Anthropic directly rather than through this repository.

Reporting a Vulnerability

Please report suspected vulnerabilities privately to accountable-review-security@wyeworks.com rather than opening a public issue.

Include, where possible:

  • The affected file(s) or generated artifact (e.g. the rendered workflow YAML).
  • Plugin version (.claude-plugin/plugin.json) or commit SHA.
  • Steps to reproduce, and the impact you believe it has (e.g. privilege escalation in CI, token exposure, arbitrary code execution).

You can expect an acknowledgment within 5 business days. We aim to provide an initial assessment (accepted, needs more information, or declined) within 14 days, and to ship a fix or mitigation for an accepted report within 30 days, depending on severity. We will credit reporters in the release notes unless you ask to remain anonymous.

If a report is declined, we will explain why.

There aren't any published security advisories