electronegativity

Project overview

SkillWard triages Agent Skills in three stages: rule-based static analysis, LLM-assisted intent evaluation, and isolated Docker execution for unresolved cases. The runtime guard records evidence of suspicious network access, sensitive writes, credential risks, and other behavior. It supports single or batch CLI scans, with an optional web interface for reviewing results.

Project overview

Place the scanner between untrusted content, model output, and tool execution, or wrap a line-delimited stdio MCP server with its proxy. It detects prompt injection, credential disclosure, exfiltration, and destructive actions before forwarding a call, redacts secrets, and returns structured reasons for enforcement. The scanner makes no network calls and ships as a Rust CLI with Node and Python wrappers.