Source confirms an automatic, opt-out project-local installation of first-party Claude Code skills, which presents residual agent extension lifecycle risk. The inspected behavior does not establish a concrete attack warranting a block.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
Installing the package outside CI when AX_SKIP_SKILL_INSTALL is not set.
Impact
The skills become available to Claude Code in the project without a separate setup command or affirmative consent prompt.
Mechanism
The postinstall script copies packaged Markdown skill files into .claude/skills, installing missing files and upgrading older versions.
Evidence
package.jsonscripts/postinstall.mjscli/index.mjsskills/*.md.claude/skills