AI Security Review
scanned 8d ago · by lpm-firewall-aiReview flagged AI-agent configuration or capability changes. This remains warn-only unless evidence shows foreign-agent hijack through preinstall/install/postinstall, hidden persistence, exfiltration, remote code execution, or other concrete malicious behavior.
Decision evidence
public snapshot- dist/chunk-SYFYHKKI.js implements <system> using child_process exec/spawn with shell/background support.
- dist/chunk-SYFYHKKI.js implements <eval> via AsyncFunction with fs/path/session context.
- lib/ai.di defines an ai helper that calls <llm execute="true">, allowing LLM-generated Dirac code execution.
- dist/agent-AQDM6UOR.js user-invoked agent start creates ~/.dirac pid/log files and a detached daemon.
- package.json has no install/preinstall/postinstall lifecycle hooks.
- dist/cli.js only activates shell, agent, or file execution from explicit CLI commands.
- No code writes Claude/Codex/Cursor/MCP or other foreign AI-agent control surfaces.
- Network calls are LLM/embedding provider features using configured OpenAI/Anthropic/local/custom endpoints.
Source & flagged code
5 flagged · loading sourceThis package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
dist/chunk-SYFYHKKI.jsView on unpkgPackage source references child process execution.
dist/chunk-SYFYHKKI.jsView on unpkg · L1617Package source references dynamic require/import behavior.
dist/chunk-SYFYHKKI.jsView on unpkg · L299Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.
dist/shell-IYJJWIQR.jsView on unpkg · L20