The source shows an automatic dependency postinstall that writes a Claude skill into the consuming project’s agent configuration directory. This is an agent extension lifecycle risk, but the inspected behavior is scoped to the dependency’s skill files and does not establish malicious activity.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review
Trigger
Installation of @ax-llm/ax as a dependency; its postinstall runs unless its skip conditions apply.
Impact
Project-local agent configuration is changed without a separate explicit command, creating an agent extension lifecycle risk.
Mechanism
The dependency copies bundled skill Markdown into the project’s .claude/skills directory.
Evidence
package.json.lpm-dependency-evidence-94a58d70-f99b-4460-bd03-9ec9eb4684f9/0/npm-linux/package.json.lpm-dependency-evidence-94a58d70-f99b-4460-bd03-9ec9eb4684f9/0/npm-linux/scripts/postinstall.mjs.claude/skills/<skill-name>/SKILL.md