The postinstall hook automatically registers the package’s own skill with multiple agent hosts, which is an agent extension lifecycle risk. The behavior is disclosed and aligned with the package’s stated purpose, so this warrants a warning rather than a block.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
Package installation, except when npm runs it through `exec`.
Impact
The skill becomes available to those agents without a separate user-run setup command; this is an agent extension lifecycle risk. The README discloses the behavior.
Mechanism
The hook launches `plugin.mjs install`, which stages the package and registers its skill with the default hosts.
Evidence
bin/postinstall.mjsbin/plugin.mjslib/install.mjsREADME.mdskills/kling-ai-plugin/SKILL.md~/.agents/skills/kling-ai-plugin~/.claude/skills/kling-ai-plugin~/.pi/agent/skills/kling-ai-plugin~/.workbuddy/skills/kling-ai-plugin~/.workbuddy-ai/skills/kling-ai-plugin~/.qwenworkcn/skills/kling-ai-plugin