The MCP entrypoint automatically writes the package’s own hooks and behavior instructions into existing agent configurations during initialization. The feature and service are documented, and the inspected content flow serves the package’s stated purpose, so this warrants a warning for automatic agent configuration changes rather than a malicious verdict.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
MCP initialize, unless SLASHML_DISABLE_AUTO_HOOKS=1; hook configuration is applied to existing agent directories.
Impact
Agent settings and instructions change persistently, and agent event content can be processed by the package’s service through the installed hooks.
Mechanism
Writes a hook launcher and user-level agent hooks and instructions. The hooks process agent events and send prompt/tool content to the package’s memory and guardrail API.
Evidence
bin/slashml-mcp.jsREADME.md~/.slashml/bin/slashml-hook.js~/.claude/settings.json~/.claude/CLAUDE.md~/.cursor/hooks.json~/.cursor/rules/slashml-memory.mdc~/.codeium/windsurf/hooks.json~/.codeium/windsurf/memories/global_rules.md