If you discover a security vulnerability in this repository, please report it privately. Do not open a public GitHub issue.
The fastest way to report is via GitHub's private vulnerability reporting:
- Go to the Security tab of this repository
- Click Report a vulnerability
- Fill out the form with as much detail as possible
This keeps the report confidential while we investigate and patch.
You can also email the security team directly at security@rampstack.co.
When reporting, please include:
- A description of the vulnerability
- Steps to reproduce
- The potential impact
- Any suggested mitigation, if you have one
- Whether you would like public credit when the fix ships
- Acknowledgment within 3 business days
- Initial assessment within 7 business days, including a severity classification
- Status updates every 7 days while the fix is in progress
- Public disclosure coordinated with the reporter once a fix is available
This repository is a library of Claude Skills (markdown files and reference documents). It contains no executable code, no servers, and no runtime. The most likely security concerns are:
- Prompt injection vectors in skill content that could manipulate an LLM consuming the skills against user intent
- Sensitive information leakage in skill content (private credentials, internal URLs, personal data)
- Malicious links in references or examples
- Misleading security guidance in skills like
security-baselineorcode-review-webthat could cause downstream harm if followed
If you find any of the above, please report it via the channels above.
The following are not security vulnerabilities for this repository:
- Issues with third-party tools recommended in skills (Ahrefs MCP, GitHub MCP, etc.). Report those to the respective vendors.
- Issues with how Claude itself handles skill content. Report those to Anthropic.
- General feedback or suggestions on skill quality. Use Issues for those.
We thank security researchers who responsibly disclose vulnerabilities. With permission, we will list contributors who help keep this library safe in this section.
No reports yet.
Thank you for helping keep this project and its users safe.