The package automatically installs its own agent skill into detected agent directories during supported global installs, which is an install-time extension lifecycle risk. It also sends runtime and error telemetry to the product endpoint without a located opt-out, so downgrade to a warning rather than block publication.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review; previous stored version diff introduced dangerous source
Trigger
Supported global package installation; runtime commands and startup errors trigger telemetry.
Impact
Agent skill directories are modified during installation, and diagnostic or invocation records leave the machine.
Mechanism
The postinstall handler synchronizes the bundled skill to canonical and detected agent skill directories. Telemetry events are posted to the configured product endpoint.
Evidence
package.jsondist/bin/postinstall.jsdist/src/features/skill/installer.jsdist/src/telemetry/sls-sink.jsdist/src/product/config.js~/.agents/skills/insmind-cli~/.claude/skills/insmind-cli~/.codex/skills/insmind-cli