@kyh/skills@0.1.3

High-risk behavior combination matched malicious policy.

`npm install -g @kyh/skills` runs the `postinstall` hook.

Changes instructions and capabilities available to multiple AI-agent tools and imports code/content from 14 external repositories without a separate user command.

Global symlink/copy injection plus automatic external skill installation.

On global installation, the postinstall hook links or copies this package’s agent instructions and skills into shared `~/.agents` and `~/.claude` locations read by several agent products. It then automatically runs the `skills` CLI (or `npx`) to install every hard-coded external repository globally. Although local installs and CI are guarded and the current MCP list is empty, this is an unconsented lifecycle-time mutation of a broad AI-agent control surface.

Source inspection confirms the scanner’s core finding: the package performs global AI-agent configuration and third-party skill installation during npm postinstall. That concrete behavior meets the blocking boundary for lifecycle mutation of a broad agent control surface. Product guard normalized a concrete AI-agent control hijack publish_block to the blockable dangerous-capability shape.