github-skills-manager

[Skill Scanner] Installation of third-party script detected No direct malicious code is present in the provided SKILL.md text, but the skill’s scope and claims raise supply-chain and confidentiality risks. The dashboard's ability to run npm install across many projects and to create/commit/push across repositories is legitimate for a manager tool but also increases attack surface: npm lifecycle scripts and third-party templates (e.g., 'skill-creator') can execute arbitrary code, and broad git operations can accidentally push secrets. The explicit mention of integrating Confidential and Personal knowledge tiers without describing safe handling is a red flag. Recommend code-level review of the scripts (dashboard.cjs, create_skill.cjs, any 'skill-creator' source), auditing for where downloads originate, ensuring least-privilege, and adding explicit checks that secrets are not staged/pushed. Treat this as SUSPICIOUS until implementation-level review and safe-guarding are confirmed. LLM verification: The project is a legitimate management tool with necessary filesystem, git, and network interactions. However, the documentation reveals high-risk capabilities (automatic integration of confidential/personal knowledge tiers, mass-install and mass-push operations, and invocation of npm install) without describing safeguards. In absence of the referenced script implementations, treat the package as potentially dangerous until a code-level review is performed. Prioritize auditing scripts for child_