Only the main branch is actively maintained (rolling release). No versioned releases exist.
This repository contains Markdown skill instructions, not executable code. The following qualify as security issues:
- Prompt injections in
SKILL.mdthat could cause Claude to take unintended actions - Accidental exposure of private data in examples (emails, tokens, internal company names)
- Malicious or misleading links in documentation
- Instructions that violate Anthropic's Usage Policies
The following are not security issues — please use a regular bug report instead:
- Typos or unclear wording
- Skill logic errors or suboptimal outputs
Preferred: Use GitHub's private vulnerability reporting — go to Security → Advisories → Report a vulnerability in this repository. This keeps the report private until resolved.
Alternative: Open a thread in GitHub Discussions and mention @KirKruglov if confidentiality is not a concern.
Reports are handled on a best-effort basis. There is no guaranteed response SLA.
After a reported issue is resolved, a public Security Advisory will be published in this repository.