The install hook automatically writes this package’s bundled skills into a consuming project’s Claude instructions, which creates an agent-extension lifecycle risk. Its scope is limited to those skills and it skips CI and non-checkout contexts, so the finding is warning-level rather than evidence of credential theft or broader agent hijacking.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review; source fingerprint signature matched known malicious package; routed for review; previous stored version diff introduced dangerous source
Trigger
Dependency installation when the package is installed in node_modules and an app is found in a Git checkout; CI and QORE_SKILLS_SYNC=0 skip the action.
Impact
The installed package can change which instructions Claude sees in the consuming project. The bundled material is package-specific skills, so this is an agent-extension lifecycle risk.
Mechanism
The hook plans skill updates and applies them without a separate sync command, writing and deleting managed files under .claude and potentially updating CLAUDE.md.
Evidence
package.jsonpostinstall.mjsdist/cli/skills-postinstall.jsskills/catalog.json.claude/**CLAUDE.md