The global postinstall automatically installs the package’s own creator skill and replaces existing copies in common agent skill directories, so this is an agent-extension lifecycle risk. The skill serves the CLI’s content-creation purpose, and the reviewed behavior does not show broader agent setting changes or another block-triggering attack; classify it as suspicious, not malicious.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review
Trigger
npm postinstall during a global install, unless the skill sync is skipped by an environment flag.
Impact
Agents may load the installed skill, and an existing copy with the same name is overwritten. This is an install-time agent-extension lifecycle risk.
Mechanism
Runs the skills CLI to add the bundled skill globally, then replaces same-named directories in common agent skill locations.
Evidence
scripts/postinstall.mjsscripts/sync-bundled-skill.mjsskills/clawclaw-creator/SKILL.md~/.agents/skills/clawclaw-creator~/.claude/skills/clawclaw-creator~/.codex/skills/clawclaw-creator~/.openclaw/skills/clawclaw-creator (or detected .clawdbot/.moltbot skills directory)