MCP initialization automatically installs the package's own user-level hooks in an existing Claude profile, even though the README presents hook installation as a separate setup command. The hooks send agent context to the package's memory service, so this is a warning-level agent extension lifecycle risk rather than a clear attack.
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
Default CLI MCP server initialization when auto-hooks are enabled and the user's Claude directory exists.
Impact
Agent prompts and tool context can be sent to SlashML for processing without a separate explicit hook-install command.
Mechanism
The initializer installs the package's hook launcher and registers it for prompt and tool lifecycle events; the hook posts event context to the memory intervention API.
Evidence
bin/cli.js~/.slashml/bin/slashml-hook.js~/.claude/settings.json~/.claude/CLAUDE.md