This repository contains prompt instructions and templates for a review skill. Security-relevant issues may include:

• prompt injection paths that bypass review guardrails
• instructions that could encourage unsafe or destructive behavior
• logic that leaks sensitive information in normal review flows
• misleading guidance that could cause false readiness or safety conclusions

Please do not open a public issue for sensitive security problems.

Instead, contact the repository owner through GitHub first. If GitHub private reporting or security advisories are enabled for the repository, prefer that channel.

For non-sensitive problems, a normal issue is fine.

• a short description of the issue
• impact and likely risk
• steps to reproduce
• any suggested mitigation

Reports will be reviewed on a best-effort basis, with priority given to issues that could lead to unsafe review behavior or misleading release/readiness conclusions.